Datasets:

akumoth

/

peewee-issues

like 0

{ "login": "tradaniel", "id": 10319846, "node_id": "MDQ6VXNlcjEwMzE5ODQ2", "avatar_url": "https://avatars.githubusercontent.com/u/10319846?v=4", "gravatar_id": "", "url": "https://api.github.com/users/tradaniel", "html_url": "https://github.com/tradaniel", "followers_url": "https://api.github.com/users/tradaniel/followers", "following_url": "https://api.github.com/users/tradaniel/following{/other_user}", "gists_url": "https://api.github.com/users/tradaniel/gists{/gist_id}", "starred_url": "https://api.github.com/users/tradaniel/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/tradaniel/subscriptions", "organizations_url": "https://api.github.com/users/tradaniel/orgs", "repos_url": "https://api.github.com/users/tradaniel/repos", "events_url": "https://api.github.com/users/tradaniel/events{/privacy}", "received_events_url": "https://api.github.com/users/tradaniel/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You don't need a pool for multi-threaded, for what it's worth. Peewee correctly handles connections and will open a connection per thread.\n\nThe issue is related to #727 and #806 . I've reverted the change in #727 so this should resolve itself. In the meantime, make sure you are using transactions! Wrap your code in `database.atomic()` blocks where you are making changes. That should help until I create a new release that has the `commit_select` change.\n", "See 6087199320a3ad8968add5d50f6c3a047767afdd\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/814/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Oh, good deal\n", "@coleifer Was this meant to be closed without merging? Does it need improvement?\n", "Whoops!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/813/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/813", "html_url": "https://github.com/coleifer/peewee/pull/813", "diff_url": "https://github.com/coleifer/peewee/pull/813.diff", "patch_url": "https://github.com/coleifer/peewee/pull/813.patch", "merged_at": "2016-01-11T16:42:05" }

{ "login": "csytan", "id": 16357, "node_id": "MDQ6VXNlcjE2MzU3", "avatar_url": "https://avatars.githubusercontent.com/u/16357?v=4", "gravatar_id": "", "url": "https://api.github.com/users/csytan", "html_url": "https://github.com/csytan", "followers_url": "https://api.github.com/users/csytan/followers", "following_url": "https://api.github.com/users/csytan/following{/other_user}", "gists_url": "https://api.github.com/users/csytan/gists{/gist_id}", "starred_url": "https://api.github.com/users/csytan/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/csytan/subscriptions", "organizations_url": "https://api.github.com/users/csytan/orgs", "repos_url": "https://api.github.com/users/csytan/repos", "events_url": "https://api.github.com/users/csytan/events{/privacy}", "received_events_url": "https://api.github.com/users/csytan/received_events", "type": "User", "site_admin": false }

[]

[]

[ "This doesn't seem like a peewee issue to me, as peewee doesn't know about your event loop and is designed using normal python APIs. I am not familiar with tornado/twisted so I can't really offer much more in the way of advice for debugging this.\n", "Understood. Thanks for having a look!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/812/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Added this. MySQL has always been a 2nd-class citizen since I don't use it and don't typically even have it installed. I've added it, however, so let's hope things work!\n", "Well, lots of issues... reopening to fix.\n", "Admittedly MySQL is a flawed technology in comparison with Postgres, but fwiw MariaDB is a fantastic improvement and also backwards compatible with Oracle/Percona MySQL, right down to the binlogs. \n\nFigured there would probably be a lot of failed tests. I can contribute a couple of hours next week to help get this working if needed (if so, just let me know where you'd like me to focus efforts)\n", "These appear to be the two errors:\n\n```\n======================================================================\nERROR: test_add_not_null (playhouse.tests.test_migrate.MySQLMigrationTestCase)\n----------------------------------------------------------------------\nTraceback (most recent call last):\n File \"/home/travis/build/coleifer/peewee/playhouse/tests/test_migrate.py\", line 278, in test_add_not_null\n dob=None)\n File \"/opt/python/2.7.9/lib/python2.7/unittest/case.py\", line 473, in assertRaises\n callableObj(*args, **kwargs)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 4493, in create\n inst.save(force_insert=True)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 4679, in save\n pk_from_cursor = self.insert(**field_dict).execute()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3213, in execute\n cursor = self._execute()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 2628, in _execute\n return self.database.execute_sql(sql, params, self.require_commit)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3461, in execute_sql\n self.commit()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3285, in __exit__\n reraise(new_type, new_type(*exc_args), traceback)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3454, in execute_sql\n cursor.execute(sql, params or ())\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/cursors.py\", line 146, in execute\n result = self._query(query)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/cursors.py\", line 296, in _query\n conn.query(q)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 819, in query\n self._affected_rows = self._read_query_result(unbuffered=unbuffered)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 1001, in _read_query_result\n result.read()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 1286, in read\n first_packet = self.connection._read_packet()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 965, in _read_packet\n packet.check_error()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 394, in check_error\n err.raise_mysql_exception(self._data)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/err.py\", line 120, in raise_mysql_exception\n _check_mysql_exception(errinfo)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/err.py\", line 112, in _check_mysql_exception\n raise errorclass(errno, errorvalue)\nIntegrityError: (1048, u\"Column 'dob' cannot be null\")\n======================================================================\nERROR: test_modify_not_null_foreign_key (playhouse.tests.test_migrate.MySQLMigrationTestCase)\n----------------------------------------------------------------------\nTraceback (most recent call last):\n File \"/home/travis/build/coleifer/peewee/playhouse/tests/test_migrate.py\", line 315, in test_modify_not_null_foreign_key\n user=None)\n File \"/opt/python/2.7.9/lib/python2.7/unittest/case.py\", line 473, in assertRaises\n callableObj(*args, **kwargs)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 4493, in create\n inst.save(force_insert=True)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 4679, in save\n pk_from_cursor = self.insert(**field_dict).execute()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3213, in execute\n cursor = self._execute()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 2628, in _execute\n return self.database.execute_sql(sql, params, self.require_commit)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3461, in execute_sql\n self.commit()\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3285, in __exit__\n reraise(new_type, new_type(*exc_args), traceback)\n File \"/home/travis/build/coleifer/peewee/peewee.py\", line 3454, in execute_sql\n cursor.execute(sql, params or ())\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/cursors.py\", line 146, in execute\n result = self._query(query)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/cursors.py\", line 296, in _query\n conn.query(q)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 819, in query\n self._affected_rows = self._read_query_result(unbuffered=unbuffered)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 1001, in _read_query_result\n result.read()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 1286, in read\n first_packet = self.connection._read_packet()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 965, in _read_packet\n packet.check_error()\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/connections.py\", line 394, in check_error\n err.raise_mysql_exception(self._data)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/err.py\", line 120, in raise_mysql_exception\n _check_mysql_exception(errinfo)\n File \"/home/travis/virtualenv/python2.7.9/lib/python2.7/site-packages/pymysql/err.py\", line 112, in _check_mysql_exception\n raise errorclass(errno, errorvalue)\nIntegrityError: (1048, u\"Column 'user_id' cannot be null\")\n----------------------------------------------------------------------\n```\n", "I'm not quite sure what's going on, those tests are passing locally for me.\n", "Only had a few minutes spare to look at this, I think the problem might be because `pymysql` is raising it's own class of `IntegrityError` seen [here](https://github.com/PyMySQL/PyMySQL/blob/master/pymysql/err.py#L44).\n\nAlso note that your `assertRaises` is looking for `_exception_not_null_violation = OperationalError`, where as MySQL is raising `IntegrityError`.\n\nSo either;\n\nA) assertRaises needs to look for `peewee.OperationalError`\nB) assertRaises needs to look for `pymysql.OperationalError`\nC) There is something wrong with `migrate.add_not_null`\n\nAs for why it's working locally, it could be due to differences in MySQL version (e.g. behaviour changes between 5.0 ~ 5.5), although 5.0 doesn't really need to be supported imho (it's as old as Python 2.6 iirc)\n", "Alright, I believe this one is now fixed. Thanks for keeping me honest @foxx!\n", "Awesome! I'm curious, what was the problem with this in the end?\n\nPS) Lol @ dropping 3.2/3.3, that's what I had to do as well, it's not widely used so should be fine.\n", "Ummmm, let's see. pymysql wasn't compat with 3.2/3.3. Then you were\ncorrect, I just needed to catch both integrity/operational errors for mysql\nfor whatever reason.\n\nOn Thu, Jan 14, 2016 at 3:59 AM, Cal Leeming [email protected]\nwrote:\n\n> Awesome! I'm curious, what was the problem with this in the end?\n> \n> PS) Lol @ dropping 3.2/3.3, that's what I had to do as well, it's not\n> widely used so should be fine.\n> \n> —\n> Reply to this email directly or view it on GitHub\n> https://github.com/coleifer/peewee/issues/811#issuecomment-171592148.\n", "Doh sorry, I just realised you pushed the code to master, I would have checked the commit history myself if I had realised lol, my bad.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/811/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "alexpantyukhin", "id": 6513121, "node_id": "MDQ6VXNlcjY1MTMxMjE=", "avatar_url": "https://avatars.githubusercontent.com/u/6513121?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alexpantyukhin", "html_url": "https://github.com/alexpantyukhin", "followers_url": "https://api.github.com/users/alexpantyukhin/followers", "following_url": "https://api.github.com/users/alexpantyukhin/following{/other_user}", "gists_url": "https://api.github.com/users/alexpantyukhin/gists{/gist_id}", "starred_url": "https://api.github.com/users/alexpantyukhin/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/alexpantyukhin/subscriptions", "organizations_url": "https://api.github.com/users/alexpantyukhin/orgs", "repos_url": "https://api.github.com/users/alexpantyukhin/repos", "events_url": "https://api.github.com/users/alexpantyukhin/events{/privacy}", "received_events_url": "https://api.github.com/users/alexpantyukhin/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You can use the `PickledField`, added just now:\n\nhttps://github.com/coleifer/peewee/blob/master/playhouse/fields.py#L293-L302\n\nIf you're using a new-ish version of Postgres you can use the JSON field or the HStore field.\n\nIf you're using a new-ish version of SQLite with the JSON1 extension compiled in or loadable, you can use the JSONField from the SQLite extension module.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/810/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "josefdlange", "id": 1062835, "node_id": "MDQ6VXNlcjEwNjI4MzU=", "avatar_url": "https://avatars.githubusercontent.com/u/1062835?v=4", "gravatar_id": "", "url": "https://api.github.com/users/josefdlange", "html_url": "https://github.com/josefdlange", "followers_url": "https://api.github.com/users/josefdlange/followers", "following_url": "https://api.github.com/users/josefdlange/following{/other_user}", "gists_url": "https://api.github.com/users/josefdlange/gists{/gist_id}", "starred_url": "https://api.github.com/users/josefdlange/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/josefdlange/subscriptions", "organizations_url": "https://api.github.com/users/josefdlange/orgs", "repos_url": "https://api.github.com/users/josefdlange/repos", "events_url": "https://api.github.com/users/josefdlange/events{/privacy}", "received_events_url": "https://api.github.com/users/josefdlange/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Yes, its because null=False. That is normal behavior. It happens whether or not you initialize the fk with None, if there is no value. The null False tells us that a missing value is an error.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/809/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "gjedeer", "id": 315648, "node_id": "MDQ6VXNlcjMxNTY0OA==", "avatar_url": "https://avatars.githubusercontent.com/u/315648?v=4", "gravatar_id": "", "url": "https://api.github.com/users/gjedeer", "html_url": "https://github.com/gjedeer", "followers_url": "https://api.github.com/users/gjedeer/followers", "following_url": "https://api.github.com/users/gjedeer/following{/other_user}", "gists_url": "https://api.github.com/users/gjedeer/gists{/gist_id}", "starred_url": "https://api.github.com/users/gjedeer/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/gjedeer/subscriptions", "organizations_url": "https://api.github.com/users/gjedeer/orgs", "repos_url": "https://api.github.com/users/gjedeer/repos", "events_url": "https://api.github.com/users/gjedeer/events{/privacy}", "received_events_url": "https://api.github.com/users/gjedeer/received_events", "type": "User", "site_admin": false }

[]

[]

[ "There is no object matching your query. Use .get() instead if you only need 1 object.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/808/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "vincentdavis", "id": 232564, "node_id": "MDQ6VXNlcjIzMjU2NA==", "avatar_url": "https://avatars.githubusercontent.com/u/232564?v=4", "gravatar_id": "", "url": "https://api.github.com/users/vincentdavis", "html_url": "https://github.com/vincentdavis", "followers_url": "https://api.github.com/users/vincentdavis/followers", "following_url": "https://api.github.com/users/vincentdavis/following{/other_user}", "gists_url": "https://api.github.com/users/vincentdavis/gists{/gist_id}", "starred_url": "https://api.github.com/users/vincentdavis/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/vincentdavis/subscriptions", "organizations_url": "https://api.github.com/users/vincentdavis/orgs", "repos_url": "https://api.github.com/users/vincentdavis/repos", "events_url": "https://api.github.com/users/vincentdavis/events{/privacy}", "received_events_url": "https://api.github.com/users/vincentdavis/received_events", "type": "User", "site_admin": false }

[]

[]

[ "How did you create the table, as in what method? You should have either called db.create_tables or model.create_table, but not db.create_table as this will not create indexes.\n", "I used db.create_table.\nThanks\n", "Yeah, it's an older API and can unfortunately cause some confusion... db.create_table() just creates the table. However create_tables() will create the tables for a number of models, resolving the foreign key dependencies to ensure a proper order of operations. Model.create_table() will just create the table and indexes for the model.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/807/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "dista", "id": 674913, "node_id": "MDQ6VXNlcjY3NDkxMw==", "avatar_url": "https://avatars.githubusercontent.com/u/674913?v=4", "gravatar_id": "", "url": "https://api.github.com/users/dista", "html_url": "https://github.com/dista", "followers_url": "https://api.github.com/users/dista/followers", "following_url": "https://api.github.com/users/dista/following{/other_user}", "gists_url": "https://api.github.com/users/dista/gists{/gist_id}", "starred_url": "https://api.github.com/users/dista/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/dista/subscriptions", "organizations_url": "https://api.github.com/users/dista/orgs", "repos_url": "https://api.github.com/users/dista/repos", "events_url": "https://api.github.com/users/dista/events{/privacy}", "received_events_url": "https://api.github.com/users/dista/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I turn on mysql log, it seems the query really go to mysql\n\n```\nmysql> select * from mysql.general_log;\n+---------------------+------------------------------------+-----------+-----------+--------------+--------------------------------------------------------------------------------------------------------------------------------------------------------+\n| event_time | user_host | thread_id | server_id | command_type | argument |\n+---------------------+------------------------------------+-----------+-----------+--------------+--------------------------------------------------------------------------------------------------------------------------------------------------------+\n| 2016-01-05 19:58:04 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log |\n| 2016-01-05 19:58:15 | tvie[tvie] @ localhost [127.0.0.1] | 83 | 0 | Query | SELECT `t1`.`id`, `t1`.`hostname`, `t1`.`processname`, `t1`.`version`, `t1`.`start_time` FROM `processstate` AS t1 WHERE (`t1`.`hostname` = 'dista81') |\n| 2016-01-05 19:58:21 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log |\n| 2016-01-05 19:58:32 | tvie[tvie] @ localhost [127.0.0.1] | 83 | 0 | Query | SELECT `t1`.`id`, `t1`.`hostname`, `t1`.`processname`, `t1`.`version`, `t1`.`start_time` FROM `processstate` AS t1 WHERE (`t1`.`hostname` = 'dista81') |\n| 2016-01-05 19:58:34 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log |\n| 2016-01-05 19:58:50 | tvie[tvie] @ localhost [127.0.0.1] | 83 | 0 | Query | SELECT `t1`.`id`, `t1`.`hostname`, `t1`.`processname`, `t1`.`version`, `t1`.`start_time` FROM `processstate` AS t1 WHERE (`t1`.`hostname` = 'dista81') |\n| 2016-01-05 19:58:51 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log |\n| 2016-01-05 19:59:43 | tvie[tvie] @ localhost [127.0.0.1] | 83 | 0 | Query | SELECT `t1`.`id`, `t1`.`hostname`, `t1`.`processname`, `t1`.`version`, `t1`.`start_time` FROM `processstate` AS t1 WHERE (`t1`.`hostname` = 'dista81') |\n| 2016-01-05 19:59:48 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log |\n+---------------------+------------------------------------+-----------+-----------+--------------+--------------------------------------------------------------------------------------------------------------------------------------------------------+\n9 rows in set (0.00 sec)\n\n```\n", "After change code to the following, it works\n\n```\n processes_q = ProcessState.select().where(ProcessState.hostname == \"dista81\")\n\n for item in processes_q:\n print item.start_time\n\n db.commit()\n```\n\nBut I do not want to write that because it is ugly, Is there any other way to make sure it will read the latest result?\n\nmy connection statement\n\n```\nconnect('mysql://xx:[email protected]:3306/lava_log_cluster')\n```\n", "I check mysql log again, I found this:\n\n```\n+---------------------+------------------------------------+-----------+-----------+--------------+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------+\n| 2016-01-05 20:56:28 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log order by event_time desc limit 10 |\n| 2016-01-05 20:56:12 | tvie[tvie] @ localhost [] | 31 | 0 | Query | select * from mysql.general_log order by event_time desc limit 10 |\n| 2016-01-05 20:56:08 | [tvie] @ localhost [127.0.0.1] | 94 | 0 | Connect | tvie@localhost on lava_log_cluster |\n| 2016-01-05 20:56:08 | tvie[tvie] @ localhost [127.0.0.1] | 94 | 0 | Query | SET NAMES utf8 |\n| 2016-01-05 20:56:08 | tvie[tvie] @ localhost [127.0.0.1] | 94 | 0 | Query | set autocommit=0 \n```\n\nthe autocommit mode in the log is 0.\n\nBut I do set autocommit=True in code:\n\n```\nrealdb = MySQLDatabase(\"lava_log_cluster\", host=\"127.0.0.1\",port=3306,user=\"xx\",passwd=\"xxx\", autocommit=True)\n#realdb = connect('mysql://xx:[email protected]:3306/lava_log_cluster')\ndb.initialize(realdb)\nprint db.get_autocommit()\ndb.set_autocommit(True)\n\n```\n\nWhy is that?\n", "We've just noticed the same issue. We've reverted to 2.6.4 which works fine, the bug was introduced in version 2.7.0.\n", "issue #727 introduced this bug, commit is required after select for mysql.\n\nWell, at least for long living connections. IF you disconnect after each request there should be no problem. Or commit at the end of each request.\n", "What about wrapping your code in a transaction?\n\n``` python\nwith db.atomic():\n processes_q = ProcessState.select().where(ProcessState.hostname == \"dista81\")\n\n for item in processes_q:\n print item.start_time\n```\n\nFurthermore, for a web app you really should not be using a single static connection. Did you read the docs here:\n\nhttp://docs.peewee-orm.com/en/latest/peewee/database.html#adding-request-hooks\n\nIf you read that and still are sure that a long-lived conn is what you want, I'd suggest wrapping each view in a transaction using the `atomic()` decorator or context manager.\n", "I use bottle and I read the document the one you specified, but I am a little confused,\n\n```\n# app.py\nfrom bottle import hook #, route, etc, etc.\nfrom peewee import *\n\ndb = SqliteDatabase('my-bottle-app.db')\n\n@hook('before_request')\ndef _connect_db():\n db.connect()\n\n@hook('after_request')\ndef _close_db():\n if not db.is_closed():\n db.close()\n\n# Rest of your bottle app goes here.\n```\n\nIf I run bottle in multi thread server like `cherrypy cherrypy Multi-threaded and very stable`, can it work?\n\nfor example:\nthread 1 call the db.connect hook, then execute the request, before thread 1 call db.close hook,\nthread 2 call db.connect hook. \n", "@goir , thanks, I manually set commit_select = True to db, and the problem is gone\n", "Database connections are thread local and are safe in multithreaded environments.\n", "thanks\n", "Wanted to get some clarification on this, since while debugging performance of my web app, I noticed that a lot of time is being spend doing commits after every SELECT (notably on requests that involve a lot of SELECT queries), which led me to finding the `commit_select = True` attribute in `MySQLDatabase`.\r\n\r\nIs my understanding correct, that if I properly wrap my request start/end with db connect/close (in my case I am just using `playhouse.flask_utils.FlaskDB`), then I can safely set `commit_select = False`?", "> Is my understanding correct, that if I properly wrap my request start/end with db connect/close (in my case I am just using playhouse.flask_utils.FlaskDB), then I can safely set commit_select = False?\r\n\r\nAhh, perhaps. Better would be to wrap your database code with the `db.atomic()` decorator, as this will disable the autocommit-type behavior.", "Hmm well my application is pretty large and there are a lot of queries all over the place so it's not really practical to wrap each one with an atomic() decorator. Unless you mean to wrap the whole request itself? But that wouldn't be great either because there are several endpoints/APIs where I would still want the write operations to commit, even if an exception happens mid-request.\r\n\r\nAlso I'm curious about this from a high-level/architecture perspective. At least in the case of web services, I would think that the expected/default behavior should be for all queries within a request to operate in the same \"commit lifecycle\" (not sure on the correct terminology here). E.g. if I am re-reading rows from the same table, I don't want the data to change mid-request if a write in a separate request changed the data.\r\n\r\nAlso, even with `commit_select = False`, I still see `set autocommit=0` at the beginning of each request in the mysql log. Is that expected?\r\n", "I am not suggesting you wrap each query with `atomic()`. That's silly.\r\n\r\nWrapping the whole request in a transaction is a very good idea and very sensible. Wrapping even distinct subsets of operations in `atomic()` is fine.\r\n\r\n> E.g. if I am re-reading rows from the same table, I don't want the data to change mid-request if a write in a separate request changed the data.\r\n\r\nThis is called isolation, and you need to wrap your blocks with `atomic()` to get this.\r\n\r\n> I still see set autocommit=0 at the beginning of each request in the mysql log. Is that expected?\r\n\r\nYes. Peewee handles all this stuff.\r\n\r\n-------------------\r\n\r\nSo to step back. Peewee operates in autocommit mode by default. No, not your database server's \"autocommit\", but it is effectively the same. This is to help beginners and not be confusing...e.g., when you save a model or call `create()`, then the operation happens and is committed immediately. But it is not ideal from a performance or even correctness point-of-view.\r\n\r\nTo achieve better performance, isolation, etc., you should make use of the `atomic()` helpers provided by Peewee (which can even be nested, committed/rolled-back mid-way through, and all sorts of stuff). For web apps, a transaction-per-request is usually a good way to go. If you need to commit midway through, you can do this:\r\n\r\n```python\r\n\r\nwith db.atomic() as txn:\r\n # do some stuff.\r\n txn.commit() # everything up til now is committed, new txn starts.\r\n\r\n # do some stuff\r\n\r\n# txn is committed implicitly at end of \"with\" block, unless\r\n# an exception occurred, in which case it is rolled-back.\r\n```\r\n\r\nI'd suggest you probably would want to read up on ACID and how COMMIT works in your db. ", "> I am not suggesting you wrap each query with `atomic()`. That's silly.\r\n\r\nOk, that's what I thought :), but just checking.\r\n\r\nThanks for the details. I do have a basic understanding of ACID and I do currently use `db.atomic`/rollback/etc in some places with expected results, but I admit my knowledge on it could be improved, especially as to exactly what's going on at the lower levels. I'll do some more reading on that. And yeah, it hadn't occurred to me that it makes sense to have each request operate in its own transaction (and make use of nested transaction and commits). I think that's the missing piece that will solve a lot of my questions/issues.\r\n\r\nThanks again for the support/explanation." ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/806/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "JoanSoConfused", "id": 16522559, "node_id": "MDQ6VXNlcjE2NTIyNTU5", "avatar_url": "https://avatars.githubusercontent.com/u/16522559?v=4", "gravatar_id": "", "url": "https://api.github.com/users/JoanSoConfused", "html_url": "https://github.com/JoanSoConfused", "followers_url": "https://api.github.com/users/JoanSoConfused/followers", "following_url": "https://api.github.com/users/JoanSoConfused/following{/other_user}", "gists_url": "https://api.github.com/users/JoanSoConfused/gists{/gist_id}", "starred_url": "https://api.github.com/users/JoanSoConfused/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/JoanSoConfused/subscriptions", "organizations_url": "https://api.github.com/users/JoanSoConfused/orgs", "repos_url": "https://api.github.com/users/JoanSoConfused/repos", "events_url": "https://api.github.com/users/JoanSoConfused/events{/privacy}", "received_events_url": "https://api.github.com/users/JoanSoConfused/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You don't need the `force_insert=True` since you are using an auto-incrementing primary key. That might be the first problem.\n\n> When I inspect the person table I see the id field that peewee added does not have autoincrement!\n\nSQLite and autoincrement is more involved than you might think. There's a reason peewee doesn't specify `AUTOINCREMENT` with SQLite, described here:\n\nhttps://www.sqlite.org/autoinc.html\n\nPeewee instead will use the unique `rowid` which uses the algorithm MAX(rowid) + 1 to generate new rowids, rather than ensuring they are unique across deletions, etc.\n", "I remove the force_insert=True and it does work! Many thank yous!\n\nI think there are still bugs here even if some of them are just bugs with the documentations.\n1. I do not see why using force_insert=True should cause the code to fail with such an unobvious error. The error should be more obvious or force_insert=True should be ok to use all the time. I was doing an insert so being able to force it is something I should be able to do, no? The documentations say \"If force_insert=True an INSERT will be issued regardless of whether or not the primary key exists.\" And I do not see the documentations warning me that I should not use it or not use it if I am using SQLite! So I think the documentations are very misleading and maybe even wrong in this case.\n2. The reason for there being no autoincrement for the SQLite would be good note in documentation so we know what going on! It is very confusing when the documentation say \"Because we have not specified a primary key, peewee will automatically add an auto-incrementing integer primary key field named id.\" but there is no warning that SQLite has special handlings!\n\nRight now the documentations are very misleading and the actual behaviors are very surprising. The documentations say it is ok to do things that it is not ok to do apparently, like use force_insert=True. They also do not explain very unexpected behaviors with SQLite like there being no autoincrement.\n\nI would contributes the documentation changes by my written English is not so good! But the documentations should be there to help us not confuse us with things that are misleading or wrong or if they do not say we should beware of surprises that would surprise us.\n\nI think it be very helpful if the documentations are fixed to be clear about how SQLite is different from what would be expected by long time SQLite users and SQLite users who have used other ORM with SQLite that works in ways different from peewee.\n", "1. Why are you calling `force_insert=True`? The docs indicate that parameter is only needed if you have a non-auto-incrementing primary key, or if you explicitly want to set the primary key to the value in the model instance.\n2. I don't see how I can document every little quirk of each db engine. Know your database, simple as that.\n\nI disagree that the docs are misleading. peewee has been around for 5 years and nobody's been that confused by the way `force_insert` works.\n", "1. I was a user of force_insert=True because I was creating an object for the first time and when I was looked at the table that peewee itself created with its default id primary key column it was like `\"id\" INTEGER NOT NULL PRIMARY KEY` with no auto-increment! So I did exactly what you and the documentation said to do: used force_insert=True because when I look at table schema it has a, as you said, \"a non-auto-incrementing primary key\"! I see not auto-increment in the column definition so it is by definition not non-auto-incrementing primary key! Now I know this wrong apparently in peewee land but the documentations are not clear at all so I ask for clarification.\n2. This is not \"little quirk\". It is peewee specific. In other ORM and hand written code they and we just use SQLite autoincrement directly. The deceiving part is peewee documents saying \"Because we have not specified a primary key, peewee will automatically add an auto-incrementing integer primary key field named id.\" but then on SQLite not actually using the built in autoincrement and instead doing its own peewee specific magic. I'm not saying peewee approach is unjustified but it is unusual and that is why it should be documented so that long time SQLite users who know how SQLite is typically work know that peewee is doing things so differently! Primary keys are not a \"little\" part of relational databases. Primary keys are very important!\n\nI must state that this my first experiences with peewee and I not impressed! I do not think I being unreasonable for asking for documentations to be clearer. And I probably am not first person to experience this. There probably others but I am first to speak up. Other people probably do what I am thinking of doing and moving to other Python ORM that has documentations that are clear and not misleading like peewee documentations current are. I want to be a user of peewee so bad but peewee is supposed to save me time and me having to deal with peewee-specific quirk only wastes time! I do not understand why you are not willing to add but like 2 sentences to documentations to help people now and in future. peewee work in unusual way and documentation should say so!\n", "Peewee does create an auto-incrementing primary key field named \"id\" automatically. In order to save CPU and RAM with SQLite, Peewee uses the native `rowid` field rather than specifically requiring globally unique auto-incrementing ids with SQLite. The docs are correct, you just misinterpreted the behavior of SQLite and Peewee in this case. I could care less which ORM you use, peewee would have saved you time if you had just used it like it is used in the docs.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/805/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "+1!\n\n> This could also open the door for potentially cleaning up Playhouse, allowing each of the extensions to be segregated into it's own plugin, keeping the core repo clean and lean.\n\nWould be a really really nice move :)\n", "Also, see http://www.devork.be/talks/pluggy/pluggy.html about pluggy.py (which lacks proper doc), the plugin facility of pytest.\n", "I'm not familiar with the setup.py magic, but that's definitely pretty cool. I recall Flask used to namespace extensions under `flask_ext` using some kind of magic, but then they moved away from it. From a personal standpoint, I'm not sure what it gets you beyond having your `import` statement slightly different. Am I missing something?\n", "In my particular use case, the primary benefits are nicer import statements and easier monkeypatching.\n\nI've got a collection of approx 15 extensions/plugins for Peewee that I'd like to open source, but some of them require a monkeypatch entry point, meaning the user has to do this;\n\n``` python\nfrom peewee_dbmanager import DBManager, monkeypatch\nmonkeypatch()\n```\n\nInstead of just this;\n\n``` python\nfrom peewee.dbmanager import DBManager\n```\n\nOr even just this;\n\n``` python\nfrom peewee import DBManager\n```\n\nDepending on how many of the additional libraries are being used, all of a sudden a single line import turns into 20 lines of boilerplate code.\n\nI've had a look at `pluggy` and the `setup.py` magic, it's quite awesome indeed but my primary concern is that this proposal would be rejected due to additional complexity, at the same time any plugin enhancement patch should cover future use cases, making `pluggy` the best candidate thus far due to maturity from `pytest`.\n", "> my primary concern is that this proposal would be rejected due to additional complexity\n\nIndeed. I'm not sure having what amounts to \"vanity\" imports would justify adding this additional complexity to peewee. Going to pass for now.\n", "I think that the \"vanity\" imports are not only less important but might be even counter productive. I mean, when reviewing someone's code, it's not immediately clear where this import comes from (is it from peewee itself? some other library?).\n\nBut on the other hand, hooks to allow extending peewee to avoid monkey patching are very important. There are many improvements that can be offered as plugins, but shouldn't go into core, that are hard to implement or maintain without proper hooks.\n\nJust today I monkey patched `Query` class, to be able to measure query execution time: https://github.com/getredash/redash/blob/feature/permissions/redash/metrics/database.py#L36-L62\n\nI've also [extended the `PostgresExtDatabase` class](https://github.com/getredash/redash/blob/feature/permissions/redash/metrics/database.py#L12-L33) for similar purposes. I would preferred to offer this as a library for all `Database` derived classes, but I'm not sure if it's possible without monkey patching without proper hooks.\n", "If the primary reason for rejection is complexity of `pluggy`, then we can achieve the same goal with just 10 lines of code added to the core, by using `entry_points` as seen in the thorough example [here](http://stackoverflow.com/questions/774824/explain-python-entry-points). This would simply look for any advertised objects, which should be callable, on `peewee_plugins` and then call them. \n\nThis would give us simple plugins with monkeypatching support, whilst also keeping the implementation clean and simple. Thoughts @coleifer / @arikfr ?\n", "@coleifer Has this been re-opened for consideration? If so, is there's anything I can do to help? Should I send a PR for POC?\n", "I reopened the issue because it seems that there may be some way to add hooks to avoid 3rd party packages needing to monkey-patch. I'd like to hear some use-cases that you all are currently considering, and that would require monkey-patching, so we can come up with a better solution.\n", "Well, I haven't really heard any use-cases where monkey-patching is currently necessary but a plugin approach would help avoid this... Please chime in if you think of something.\n", "There is the metrics tracking use case I had:\nhttps://github.com/getredash/redash/blob/feature/permissions/redash/metrics/database.py#L38-L64\n\nOr being able to override `execute_sql` of the base `Database` class: \nhttps://github.com/getredash/redash/blob/feature/permissions/redash/metrics/database.py#L18-L29\n\nI would love to generalize these two into a plugin.\n", "In the first case extending via subclass + override seems correct. In the second case I'd hardly call that a common use-case...\n", "I've tried subclassing first -- I've subclassed `Model` and changed all the query methods to return wrapped query object, but then it was lost when `.clone` was called on the query object. Obviously I can override it too, but it becomes more an more complex, and I ended up monkey patching the methods on the base class.\n\nOn the other hand if there were `before_execute` and `after_execute` hooks, I could implement it easily without overriding or monkey patching.\n\nThe second case is part of the same thing I tried to accomplish -- measure query execution time and number of queries executed per web request. And the same thing here: simple before/after hooks on execute would be enough.\n", "For my use case, creating a database manager for Peewee requires monkeypatching to override `_execute` and `Node`, similar to @arikfr, but as he mentioned if Peewee had a hooks system then monkeypatching would not be necessary at all.\n", "After some testing and consideration, I'm not convinced that `entry_points` is the correct approach for monkeypatch hooks. If this approach was used, then installing and importing a plugin would cause monkeypatching to happen behind the scenes, breaking the mantra of \"explicit is better than implicit\", meaning it's cleaner to call `someplugin.monkeypatch()`.\n\nThe problem in mine and @arikfr's use case is the inability to change Peewee internals without modifying global namespace, and a lack of hooks such as `before_execute` and `after_execute`. However even if we had those hooks, they would still impact global namespace due to the design of Peewee. Although we can subclass `Query` with the `_execute()` modifications, then apply this by subclassing `Model` and overloading `select()`, doing so could cause future problems because the entire `select()` class needs to be duplicated and modified.\n\nOne approach would be assigning class attributes to `Model` and `Database`, which are then used to determine the target class used. This approach could be quite tedious however, resulting in a lot of boilerplate on each class. For example;\n\n``` python\nclass Model(with_metaclass(BaseModel)):\n select_query_class = SelectQuery\n\n @classmethod\n def select(cls, *selection):\n query = self.select_query_class(cls, *selection)\n\n```\n\nInstead it might be cleaner to use a single method which a subclass can implement;\n\n``` python\nclass Model(with_metaclass(BaseModel)):\n @classmethod\n def get_class(cls):\n return cls\n\n @classmethod\n def select(cls, *selection):\n query = self.get_class(SelectQuery)(cls, *selection)\n\nclass MySelectQuery(SelectQuery):\n def _execute(self, *args, **kwargs):\n print(\"do something\")\n super()._execute(*args, **kwargs)\n\nclass MyModel(Model):\n @classmethod\n def get_class(cls):\n if cls == SelectQuery:\n return MySelectQuery\n return cls\n```\n\nThe above would allow us to easily patch Peewee internals on a per instance basis, without resorting to monkeypatching or peewee core duplication.\n\nI've done some trials of implementing this into Peewee and the results so far have been fantastic, very little change in performance and no monkeypatching required.\n\nAssuming this was accepted, I would probably be -0 on adding plugins support for now, as I wouldn't have a use case for it any longer.\n\nThoughts @arikfr / @coleifer ?\n", "Urgh, yeah I'm definitely not loving it.\n", "@coleifer This is pretty much the cleanest implementation I can think of, and after trying it some more, I've almost come to the conclusion that this isn't particularly clean from a core lib standpoint either.\n\nAnother option is module inheritance [overloading](http://stackoverflow.com/questions/7483022/clone-a-module-and-make-changes-to-the-copy), although this isn't always guaranteed to work is classes have been referenced in the original import, and therefore wouldn't be advisable.\n\nThe only clear path I'm seeing is sadly monkeypatching `peewee` module on import, and using something like [this](https://gist.github.com/foxx/87e44e00901ebddb5f93) to apply late subclasses. This has been the approach I've used since day 1, and has served me well for 6 months so far.\n\nSadly this does mean that my vote in this issue would be \"wontfix\", on the grounds that there doesn't appear to be a better alternative than monkeypatching.\n\nInstead of a code modification, perhaps we could include a docs update which explains how to do monkeypatching and extending, along with an example, using the tools shown in the above mentioned [gist](https://gist.github.com/foxx/87e44e00901ebddb5f93). Thoughts?\n", "+1 for avoiding entry points for the implicitness of it (and the need to publish plugins as Python packages).\n\nHow about starting with a set of simple hooks (like the before/after execute), that are global to peewee in the context of a single codebase/application? \n\nThe implementation for the above should be straightforward and still allow many interesting extensions. If @coleifer is cool with that, I can make pull request with a suggestion for implementation.\n", "@arikfr Could you give an example of how these hooks might be implemented and exposed? Every option I've explored has resulted with unnecessary complexity in the core, for a feature that could be easily implemented with monkey patching. Having `pre_execute` and `post_execute` can be easily done with;\n\n``` python\nclass Query(peewee.Query):\n def _execute(self, *args, **kwargs):\n print(\"pre\")\n r = super()._execute(*args, **kwargs)\n print(\"post\")\n return r\n\npeewee.Query = Query\n```\n\nFrom what I can see, the only benefit of extending Peewee internals to support hooks would be if they were on a per instance basis, however due to the design of Peewee this cannot be easily achieved (for reasons mentioned earlier, least not without a lot of redesign which results in something even more convoluted).\n\nThat being said, let me know if I've overlooked anything and I'll happily reconsider/review.\n", "Thought I'd share my discoveries from #825, as getting monkeypatching to work consistently has been troublesome.\n\nThere is an outstanding [bug](http://bugs.python.org/issue672115) in Python which prevents dynamic modification of `__bases__`, meaning that classes which are an immediate subclass of `object` cannot be extended, where as classes which are a distant subclass can. For example;\n\n``` python\nclass ModelOptions(object): pass # this cannot be extended in-line\nclass SqliteDatabase(Database): pass # this can\n```\n\nFor classes that cannot be modified inline, the alternative approach is to overwrite classes in peewee namespace, before anything else has a chance to access them. However this is not always guaranteed to work, as seen here;\n\n``` python\nimport peewee\n\nclass ModifiedSqliteDatabase(peewee.SqliteDatabase):\n pass\n\npeewee.SqliteDatabase = ModifiedSqliteDatabase\npeewee.SqliteDatabase(':memory:')\n```\n\n``` python\n File \"peewee-2.8.0-py3.4.egg/peewee.py\", line 3606, in __init__\n File \"peewee-2.8.0-py3.4.egg/peewee.py\", line 3606, in __init__\n File \"peewee-2.8.0-py3.4.egg/peewee.py\", line 3606, in __init__\nRuntimeError: maximum recursion depth exceeded\n```\n\nThis also means putting your `monkeypatch()` call before anything else, something you cannot guarantee if your application is loaded by another tool, such as py.test. Any objects created prior to the monkeypatch will use the old class, hence why this \"late\" approach is not preferred.\n\nOther approaches include using a custom metaclass, such as subclassing `BaseModel` for `Model`, however this will break other libraries using the same hacks, due to [multiple inheritance inconsistencies](http://stackoverflow.com/questions/3003053/metaclass-multiple-inheritance-inconsistency).\n\nAnd finally, you can manipulate an existing class, rather than applying a mixin subclass. Naturally you'll lose the ability to do `issubclass()` or `isinstance()` comparisons against your mixin, but chances are you will not need this anyway.\n\n``` py\ndatabase = property(getter, setter, del)\npeewee.ModelOptions.database = database\n```\n\ntl;dr - Different objects will require different monkeypatching approaches, and hopefully the above will save you a couple of hours.\n", "Does the custom base ModelOptions commit in #825 (e773c620de3d7dc686dab3c2325e02d1e86407d6) address some of the concerns brought up here?\n", "Looks like #825 resolves every issue I had with implementing a database manager for Peewee. \n\nThere are still some situations where monkeypatching cannot be avoid, as adding certain hooks would add significant mess/complexity into the core, as seen in earlier discussions.\n\nI'd be happy to contribute a docs patch explaining some of the monkeypatching tricks above, if you feel it would be beneficial, although it can be argued that this is a topic outside of Peewee.\n\nOther than a possible docs patch, I'd +1 closing this issue.\n", "Oh there is something that comes to mind, relating specifically to the example from @arikfr.\n\nIn his use case, he was overloading `Query._execute()` in order to measure execution times. Doing this by overloading `Model.select()` isn't really an option, because it would require copy/paste of Peewee internals, as mentioned earlier.\n\nFor his situation, a similar patch might also help with those cases;\n\n``` python\nclass A(Model):\n class Meta:\n select_query_class = CustomSelectQuery\n raw_query_class = CustomRawQuery\n ..etc etc\n```\n\nThis is the same approach as the `ModelOptions` patch, but it does result in more mess as these classes are referenced in several places. From an implementation standpoint, it would look like this;\n\n``` python\n @classmethod\n def select(cls, *selection):\n query = self._meta.select_query_class(cls, *selection)\n if cls._meta.order_by:\n query = query.order_by(*cls._meta.order_by)\n return query\n```\n\nThat being said, this would handle a significant range of plugin functionality without the need for monkeypaching. It also prevents peewee namespace from being modified, which is a huge benefit.\n\nIf this was in the core, the only instance I can think of where monkeypatching would be needed would be hotfixing internals.\n\nAssuming @coleifer was happy with the additional mess caused, I would certainly be +1\n", "Ehh, I'm not quite sure again why you couldn't just override `select`. Is it that common of an occurance that you need a custom `SelectQuery` implementation?\n", "Actually I just realised something, you can override `insert()` without having to copy a bunch of internals as seen;\n\n``` python\n@classmethod\n def insert(cls, __data=None, **insert):\n fdict = __data or {} # needs copying\n fdict.update([(cls._meta.fields[f], insert[f]) for f in insert]) # needs copying\n return MyInsertQueryHere(cls, fdict) # custom lines here\n```\n\nBecause the `Query()` is an instance, it becomes super easy to extend `__bases__` with none of the horrible monkeypatching bugs I mentioned earlier. For example;\n\n``` python\n@classmethod\n def insert(cls, __data=None, **insert):\n r = super(Model, cls).insert(data, **insert)\n extend_instance(r, InsertQueryMixin) # see below for gist\n return r\n```\n\nYou can find my version of `extend_instance` [here](https://gist.github.com/foxx/5807755d92a76970dae6), it's fully doc tested and been used in prod with no problems.\n\nI can't think of any scenario that the above won't cover, including the use case mentioned by @arikfr. \n\nIf anyone can't make their plugin work using the variety of solutions mentioned above and in this thread, then please let me know and I'll do my best to help.\n\nThanks for the feedback @coleifer !\n", "Thanks for the comments, hopefully folks will be able to find this thread if they encounter similar issues.\n", "> Is it that common of an occurrence that you need a custom SelectQuery implementation?\n\nI also have a custom SelectQuery, and being able to just change a Meta option to use it instead of overriding `select` (including some internals copy-pasting) would be really awesome.\nSo I really appreciate the suggestion of @foxx exposed in [this comment](#issuecomment-173238786) (actually I was to suggest the same approach because I'm already [using](https://github.com/etalab/ban/blob/master/ban/db/model.py#L22) it).\n\n@coleifer would you consider a PR for this?\n", "@yohanboniface You can actually achieve this behaviour by using late mixins via [`extend_instance`](https://gist.github.com/foxx/5807755d92a76970dae6). \n\nThis function injects other classes into an instance `__bases__`, allowing for only that specific instance to be affected by the patch. Although it might feel a bit strange relying on a non standard function, this is a tried and tested technique for monkeypatching and has proven very reliable. I've actually included it as part of a helpers library I've [released](https://github.com/foxx/bottlecap/blob/master/helpful.py), but it's not yet documented.\n\n``` python\nclass SelectQueryMixin(peewee.SelectQuery):\n def _get_result_wrapper(self):\n super_func = super(SelectQueryMixin, self)._get_result_wrapper\n return getattr(self, '_result_wrapper', None) or super_func()\n\n def __len__(self):\n return self.count()\n\n\nclass Model(peewee.Model):\n class Meta:\n database = default\n\n @classmethod\n def select(cls, *args, **kwargs):\n obj = super(Model, cls).select(*args, **kwargs)\n extend_instance(obj, SelectQueryMixin)\n return obj\n```\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/804/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I've always felt that this type of thing did not belong as a part of peewee. I will pass on this feature.\n", "Got it, I'll put this into a separate repo, thanks for the quick reply!\n", "So there is no way to create a database using peewee, right?\n", "@gitttt -- just use the tool from your vendor. For sqlite you don't have to do anything, as it will create if not exists. For MySQL:\n- https://dev.mysql.com/doc/refman/5.5/en/mysqladmin.html\n- https://dev.mysql.com/doc/refman/5.5/en/create-database.html\n\nFor postgresql, use `createdb`: http://www.postgresql.org/docs/9.1/static/app-createdb.html\n", "ok, thanks\n", "for future MySQL readers, the way I chose to do this (after ending up here) was this: \r\n\r\n```\r\n db = MySQLdb.connect(host='', user='', passwd='')\r\n c = db.cursor()\r\n c.execute(\"drop database %s\" % db_name)\r\n```" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/803/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Needs tests in order to merge.\n", "You can add tests to the `test_fields.TestFieldTypes` or you can just create a one-off test-case. Either is fine.\n", "Done. Apologies for the initial lack of tests, I got lazy!\n", "Those tests don't actually test the binary field.\n", "Ah crap, it looks like my cherry pick didn't apply properly due to lack of concentration, will get this fixed in the morning, sorry\n", "Done, I'm pretty sure this is fine but any problems let me know\n", "Thank you!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/802/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/802", "html_url": "https://github.com/coleifer/peewee/pull/802", "diff_url": "https://github.com/coleifer/peewee/pull/802.diff", "patch_url": "https://github.com/coleifer/peewee/pull/802.patch", "merged_at": "2015-12-29T13:35:36" }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Thank you, merged #802.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/801/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Although I agree this is a pretty clean patch, I am going to pass on merging.\n\nI think that database routing is a complex enough topic that I do not want to include it in peewee itself (as I don't want to maintain it). One option would be to create a new package `peewee-routing` or something similar that encapsulates this logic.\n", "Thanks for the quick reply! If I put myself forward as a maintainer for this, and submitted a fully documented PR, would you consider merging into playhouse? As a last option I could create a new repo, but would argue the use cases are common enough to justify being in the playhouse.\n", "I appreciate the offer and don't want to seem to be discouraging, but I'm actively trying to rein in some of the sprawl in the code-base and tighten things up. Database-routing, to me, both complex and uninteresting from a maintenance perspective. It is my strong preference to keep it out of this repo.\n", "Okay no worries, I'll look into creating a separate repo for this. I'll update the original issue with the URL once finished, if you're happy with it then (and agree with the approach used), then perhaps it might be worthy of a small section in the docs, so at least people know where to look?\n", "> then perhaps it might be worthy of a small section in the docs, so at least people know where to look?\n\nI'm sure it'll turn out well, and I'd be happy to link it up.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/800/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/800", "html_url": "https://github.com/coleifer/peewee/pull/800", "diff_url": "https://github.com/coleifer/peewee/pull/800.diff", "patch_url": "https://github.com/coleifer/peewee/pull/800.patch", "merged_at": null }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "The only legit use case I can think of is for circular foreign keys. Which are also a code smell...\n", "Ah I hadn't considered this. There are some legit situations where you might need circular foreign keys, for example a self relating [hierarchical table](http://stackoverflow.com/a/3891612/1267398) (with most other situations being an anti pattern). Are there any better ways that circular FK dependancies could be exposed? If not, then `Proxy` deprecation is probably a no-go.\n", "Hierarchical self references are already special cases through the use of\nForeignKeyField('self'). Also many to many junction tables when used with\nManyToManyField.\n\nHonestly, we probably can ditch it and add specific APIs for the valid use\ncases.\nOn Dec 28, 2015 8:32 AM, \"Cal Leeming\" [email protected] wrote:\n\n> Ah I hadn't considered this. There are some legit situations where you\n> might need circular foreign keys, for example a self relating hierarchical\n> table http://stackoverflow.com/a/3891612/1267398. Are there any better\n> ways that circular FK dependancies could be exposed? If not, then Proxy\n> deprecation is probably a no-go.\n> \n> —\n> Reply to this email directly or view it on GitHub\n> https://github.com/coleifer/peewee/issues/799#issuecomment-167581145.\n", "Oh beautiful, I didn't know Peewee had self-referential FKs. In that case, I'm +1 for deprecation.\n", "Alright, so places where `Proxy` is used in the code-base (or tests):\n1. To allow circular foreign key dependencies. Example would be `Tweet` has a foreign key to `User` to represent the user who created the tweet, and `User` has a foreign key to `Tweet` to represent the user's favorite tweet.\n2. To allow `ManyToManyField()` to reference a custom junction model that, by necessity, has to be defined after both models. This is similar to 1.\n3. To allow users to define a database object, reference it in their code, then dynamically assign at run-time. This occurs typically when a user wants to run with one database engine for tests, one for local dev, and one for prod.\n\nTo remove `Proxy` I'd like to have suitable replacements for these use-cases. Now, none of these are impossible without Proxy, proxy just makes them convenient and is used in the docs.\n\nPriority-wise, this is very low on my list, since any replacement will necessarily require tests and docs changes, and has potential negative impact on users.\n\nI think I'm going to close this for now and will reopen when I have some time.\n", "Okay sounds good, happy to help if needed. \n\nFor item 3, I've been able to achieve this behaviour easily without `Proxy()`, so I think this can be handled with a good docs example.\n", "Getting a start, added `DeferredRelation` for foreign keys, and `DeferredThroughModel` for many-to-many.\n", "I was thinking about it and while it's a kludge, I think a lot of people have a hard time structuring their code without the Proxy object as a stand-in for their database. I've made some changes, but am going to keep proxy as the recommended way for separating the definition of model classes from a database implementation.\n", "Sounds good, at least we now have `DeferredRelation` and `DeferredThroughModel` for encouraging a better paradigm.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/799/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Fixed.\n", "Thanks :)\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/798/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "@coleifer any feedback on this? :)\n", "LGTM. I haven't tested with Postgres, but the changes looks sane. Changing from BaseModel to TestModel allows it to be ran with `--engine postgres`, and changing from TextField to BlobField allows bytes to be stored safely. Needs sign off from @coleifer \n", "Thank you!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/797/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/797", "html_url": "https://github.com/coleifer/peewee/pull/797", "diff_url": "https://github.com/coleifer/peewee/pull/797.diff", "patch_url": "https://github.com/coleifer/peewee/pull/797.patch", "merged_at": "2015-12-28T22:20:46" }

{ "login": "davidthewatson", "id": 150892, "node_id": "MDQ6VXNlcjE1MDg5Mg==", "avatar_url": "https://avatars.githubusercontent.com/u/150892?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davidthewatson", "html_url": "https://github.com/davidthewatson", "followers_url": "https://api.github.com/users/davidthewatson/followers", "following_url": "https://api.github.com/users/davidthewatson/following{/other_user}", "gists_url": "https://api.github.com/users/davidthewatson/gists{/gist_id}", "starred_url": "https://api.github.com/users/davidthewatson/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/davidthewatson/subscriptions", "organizations_url": "https://api.github.com/users/davidthewatson/orgs", "repos_url": "https://api.github.com/users/davidthewatson/repos", "events_url": "https://api.github.com/users/davidthewatson/events{/privacy}", "received_events_url": "https://api.github.com/users/davidthewatson/received_events", "type": "User", "site_admin": false }

[]

[]

[ "in case anyone else stumbles upon this combination (Heroku Postgres with peewee), the trick seems to be to parse the Heroku-provided URL first, and then provide the parsed components to the PostgresqlDatabase constructor. It looks like this:\r\n\r\n```python\r\nurl = urlparse.urlparse(os.environ[\"DATABASE_URL\"])\r\ndb = PostgresqlDatabase(database=url.path[1:], user=url.username, password=url.password, host=url.hostname, port=url.port)\r\n```\r\n\r\nSource:\r\nhttps://swifthorseman.com/2015/06/18/deploying-a-flask-app-with-peewee-to-heroku/", "Thanks @gregsadetsky " ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/796/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I opted for a slightly different route for getting the ID list. Thanks for the diff.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/795/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/795", "html_url": "https://github.com/coleifer/peewee/pull/795", "diff_url": "https://github.com/coleifer/peewee/pull/795.diff", "patch_url": "https://github.com/coleifer/peewee/pull/795.patch", "merged_at": null }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "ManyToManyField does not call `model_class.add_field`, so a \"solution\" may be to just add this to `ManyToManyField.add_to_class`:\n\n```\nmodel_class._meta.fields[name] = self\n```\n\n(It's not DRY, but we can't call `add_field` itself, because it expects the field to have a column, so maybe a bit of refactor is needed.)\n\nEdit: it crashes in `_fields_to_index` (`AttributeError: 'ManyToManyField' object has no attribute 'index'`), so certainly a filter needs to be added there.\n", "It's not technically a field, so it is not present in the field lists. It's just a descriptor that provides convenient access to the through table relationship.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/794/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "samsoft00", "id": 5300891, "node_id": "MDQ6VXNlcjUzMDA4OTE=", "avatar_url": "https://avatars.githubusercontent.com/u/5300891?v=4", "gravatar_id": "", "url": "https://api.github.com/users/samsoft00", "html_url": "https://github.com/samsoft00", "followers_url": "https://api.github.com/users/samsoft00/followers", "following_url": "https://api.github.com/users/samsoft00/following{/other_user}", "gists_url": "https://api.github.com/users/samsoft00/gists{/gist_id}", "starred_url": "https://api.github.com/users/samsoft00/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/samsoft00/subscriptions", "organizations_url": "https://api.github.com/users/samsoft00/orgs", "repos_url": "https://api.github.com/users/samsoft00/repos", "events_url": "https://api.github.com/users/samsoft00/events{/privacy}", "received_events_url": "https://api.github.com/users/samsoft00/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Wait... I'm confused by your code. You're doing a select? I think you copy/pasted wrong.\n\nThis is really not appropriate for github, please ask on stackoverflow.\n", "sorry for the mistake in updateEntries method, my intention was to delete all entries in Job table and everything that associate with it using this code \n\n```\nquery = models.Job.delete().where(models.Job.timestamp < datetime.date.today()-datetime.timedelta(days=2)) \nquery.execute()\n```\n\n but generate \n\n```\npeewee.IntegrityError: (1451, 'Cannot delete or update a parent row: a foreign key constraint fails (career.category_job_through, CONSTRAINT category_job_through_ibfk_2 FOREIGN KEY (job_id) REFERENCES job (id))')\n```\n\nwhat can i do?\n", "You need to first delete the rows from the through table that reference the job you're deleting.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/793/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "adsahay", "id": 375799, "node_id": "MDQ6VXNlcjM3NTc5OQ==", "avatar_url": "https://avatars.githubusercontent.com/u/375799?v=4", "gravatar_id": "", "url": "https://api.github.com/users/adsahay", "html_url": "https://github.com/adsahay", "followers_url": "https://api.github.com/users/adsahay/followers", "following_url": "https://api.github.com/users/adsahay/following{/other_user}", "gists_url": "https://api.github.com/users/adsahay/gists{/gist_id}", "starred_url": "https://api.github.com/users/adsahay/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/adsahay/subscriptions", "organizations_url": "https://api.github.com/users/adsahay/orgs", "repos_url": "https://api.github.com/users/adsahay/repos", "events_url": "https://api.github.com/users/adsahay/events{/privacy}", "received_events_url": "https://api.github.com/users/adsahay/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You need to probably specify a custom `RowConverter` and override the `Loader.get_converter` method to appropriately handle your CSV data.\n", "You can pass in a custom converter using the `converter` parameter.\n", "I am trying the following (no luck):\n\n```\n # use a custom RowConverter to handle boolean values, as it is not\n # handled out of the box.\n class CustomRowConverter(RowConverter):\n def field(field_class, **field_kwargs):\n def decorator(fn):\n fn.field = lambda: field_class(**field_kwargs)\n return fn\n\n return decorator\n\n @field(BooleanField, default=0)\n def is_boolean(self, value):\n # True if value is a boolean\n # ref: http://www.postgresql.org/docs/9.4/static/datatype-boolean.html\n print('is_boolean %s' % value)\n true_literals = ['TRUE', 't', 'true', 'y', 'yes', 'on', '1']\n false_literals = ['FALSE', 'f', 'false', 'n', 'no', 'off', '0']\n return value in true_literals or value in false_literals\n\n def get_checks(self):\n \"\"\"Return a list of functions to use when testing values.\"\"\"\n return [\n self.is_date, self.is_datetime, self.is_integer, self.is_float,\n self.is_boolean, self.default\n ]\n\n custom_converter = CustomRowConverter(database=database)\n try:\n load_csv(MyModel,\n 'file.csv',\n converter=custom_converter)\n except Exception as e:\n app.logger.error('Exception while loading csv %s - %s' %\n (e.__class__, e))\n```\n\nCan you please tell me where the problem is? I can't trigger the `is_boolean` check above.\n", "I put together a test-case and was able to verify that the custom converter could interpret a booleanfield. Maybe you can look at the new test-case and get some ideas on the implementation. It's a little bit of a crummy API, I definitely would like to take a crack at rewriting it in the future, but for now hopefully this helps.\n", "See: 58d0fcbb75faa6200ea63db11a4c43e304f39c73\n", "Thanks for this. I think if we supply a `model` instead of a `database` to the loader custom converter isn't working. \n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/792/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "user5642", "id": 16286595, "node_id": "MDQ6VXNlcjE2Mjg2NTk1", "avatar_url": "https://avatars.githubusercontent.com/u/16286595?v=4", "gravatar_id": "", "url": "https://api.github.com/users/user5642", "html_url": "https://github.com/user5642", "followers_url": "https://api.github.com/users/user5642/followers", "following_url": "https://api.github.com/users/user5642/following{/other_user}", "gists_url": "https://api.github.com/users/user5642/gists{/gist_id}", "starred_url": "https://api.github.com/users/user5642/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/user5642/subscriptions", "organizations_url": "https://api.github.com/users/user5642/orgs", "repos_url": "https://api.github.com/users/user5642/repos", "events_url": "https://api.github.com/users/user5642/events{/privacy}", "received_events_url": "https://api.github.com/users/user5642/received_events", "type": "User", "site_admin": false }

[]

[]

[ "SQLite does not have a bigint type natively. Check out the sqlite docs on type affinities. Regarding the specific issue it sounds to me like it's coming from the SQLite driver and not peewee.\n", "Sorry it was not the best illustration of problem encountered by me. This code should be more demonstrative:\n\n```\nfrom peewee import *\n\ndb = SqliteDatabase('test.db')\n\nclass UInt32(Field):\n db_field = 'int'\n\n def db_value(self, value):\n return value - (1<<31)\n\n def python_value(self, value):\n return value + (1<<31)\n\nclass Foo(Model):\n id = UInt32(primary_key = True)\n\n class Meta:\n database = db\n\nclass Bar(Model):\n id = UInt32(primary_key = True)\n foo = ForeignKeyField(Foo, related_name='bars')\n\n class Meta:\n database = db\n\ndb.create_tables([Foo, Bar])\n\nf = Foo.create(id=(1<<31)+1000)\nBar.create(id=0, foo=f)\nb = Bar.get()\nb.foo\n```\n\nThis will fall with following error:\n\n```\n__main__.FooDoesNotExist: Instance matching query does not exist:\nSQL: SELECT \"t1\".\"id\" FROM \"foo\" AS t1 WHERE (\"t1\".\"id\" = ?)\nPARAMS: [-2147484648]\n```\n\nIt's clear that it tries to query for entry with id equal to 1000 - (1<<31) instead of 1000. So it's seems that peewee applies `db_value` where it should not. (or other way around not applying `python_value` where it should have)\n", "It applies db_value when sending values to the database. So technically it shouldn't be necessary to do the conversion yourself when calling create(), right?\n", "\"Python id\" is equal to (1<<31)+1000 = 2147484648 (I specially chose value outside of signed int32 range for demonstration purposes)\n\"Database id\" is equal to (1<<31)+1000 - (1<<31) = 1000 (record with such id is successfully created in DB)\nBut `b.foo` for some reason makes query for id -2147484648 = 1000 - (1<<31) , not for 1000 as it should have been.\n", "Maybe try this:\n\n``` python\nclass UInt32(Field):\n db_field = 'int'\n\n def db_value(self, value):\n return long(value - (1<<31))\n\n def python_value(self, value):\n return long(value + (1<<31))\n```\n", "Actually, I think you may be onto something, I'm looking into it.\n", "Alright, I believe the commit in 9e9e0bd should have things fixed.\n", "Now it works as expected.Thank you!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/791/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Sweet, thanks!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/790/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/790", "html_url": "https://github.com/coleifer/peewee/pull/790", "diff_url": "https://github.com/coleifer/peewee/pull/790.diff", "patch_url": "https://github.com/coleifer/peewee/pull/790.patch", "merged_at": "2015-12-10T17:48:58" }

{ "login": "macks22", "id": 3409997, "node_id": "MDQ6VXNlcjM0MDk5OTc=", "avatar_url": "https://avatars.githubusercontent.com/u/3409997?v=4", "gravatar_id": "", "url": "https://api.github.com/users/macks22", "html_url": "https://github.com/macks22", "followers_url": "https://api.github.com/users/macks22/followers", "following_url": "https://api.github.com/users/macks22/following{/other_user}", "gists_url": "https://api.github.com/users/macks22/gists{/gist_id}", "starred_url": "https://api.github.com/users/macks22/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/macks22/subscriptions", "organizations_url": "https://api.github.com/users/macks22/orgs", "repos_url": "https://api.github.com/users/macks22/repos", "events_url": "https://api.github.com/users/macks22/events{/privacy}", "received_events_url": "https://api.github.com/users/macks22/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I think I will pass on this functionality.\n", "@macks22 but you know what key was passed to your function. Why not use it instead of taking this value from the exception? \n", "@arikfr that's basically what I was getting at...there's no new info there.\n", "Yes, this is what I ended up doing. I guess I just have a habit of packing as much info as possible into my exception instances. In this case it's not necessary though. Thanks for the quick feedback.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/789/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Ahh yeah good call, and I should know cause I got bitten by just this issue.\n\nYeah a PR b great, thx\n", "Awesome, PR coming shortly :)\n", "Also there could be an argument for storing as `BINARY` instead of TEXT, but I'd need to think about the edge cases. I'll open a new issue if I can think of a compelling argument\n", "@coleifer Any idea when this fix will make it into next release? Just got stung by it again lol\n", "Ahh, dang have I not pushed a release? OK I'll get one out.\n", "PS - I'm not sure if you use MySQL/MariaDB but I had to modify the UUID type to VARCHAR(255) in order to get it working. Not sure if that matters, it was LONGTEXT before.\n", "Ah sorry that's my bad, although for indexing purposes it should probably be `CHAR(16)`. I've added `OptimizedUUIDField` and `BinaryUUIDField` into my `peewee-extras` repo, for anyone that wants to store UUIDs in MySQL on a larger scale. It'll be a couple of weeks before the repo is ready for prime time, as I haven't finished adding all the tests yet.\n", "PS @foxx you may want to add a LICENSE file to peewee-extras, as I'd like to maybe borrow some of those tidbits :)\n", "@coleifer Yup will do, in a few weeks it will be as clean as [bottlecap](https://github.com/foxx/bottlecap) (travis/coveralls/py2.7/py3.4/100% test coverage etc), will ping you when ready.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/788/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "perplexedfingers", "id": 14918698, "node_id": "MDQ6VXNlcjE0OTE4Njk4", "avatar_url": "https://avatars.githubusercontent.com/u/14918698?v=4", "gravatar_id": "", "url": "https://api.github.com/users/perplexedfingers", "html_url": "https://github.com/perplexedfingers", "followers_url": "https://api.github.com/users/perplexedfingers/followers", "following_url": "https://api.github.com/users/perplexedfingers/following{/other_user}", "gists_url": "https://api.github.com/users/perplexedfingers/gists{/gist_id}", "starred_url": "https://api.github.com/users/perplexedfingers/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/perplexedfingers/subscriptions", "organizations_url": "https://api.github.com/users/perplexedfingers/orgs", "repos_url": "https://api.github.com/users/perplexedfingers/repos", "events_url": "https://api.github.com/users/perplexedfingers/events{/privacy}", "received_events_url": "https://api.github.com/users/perplexedfingers/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You have to join on permission since you're selecting from it. That will never change.\n\nIf you don't need the order by, why are you calling it? I'm confused why that's an issue.\n", "It requires an action after `join()`.\n", "I figure it out. Sorry about wasting your time.\n\n``` python\n(PermissionInGroup.select(Permission.name,\n Group.name,\n PermissionInGroup)\n .join(Permission)\n .switch(PermissionInGroup)\n .join(Group)\n .join(UserInGroup)\n .join(User)\n .where(User.ID == user_id,\n Group.app_account_id == app_account_id,\n Group.company_id == company_id)\n)\n```\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/787/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "macks22", "id": 3409997, "node_id": "MDQ6VXNlcjM0MDk5OTc=", "avatar_url": "https://avatars.githubusercontent.com/u/3409997?v=4", "gravatar_id": "", "url": "https://api.github.com/users/macks22", "html_url": "https://github.com/macks22", "followers_url": "https://api.github.com/users/macks22/followers", "following_url": "https://api.github.com/users/macks22/following{/other_user}", "gists_url": "https://api.github.com/users/macks22/gists{/gist_id}", "starred_url": "https://api.github.com/users/macks22/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/macks22/subscriptions", "organizations_url": "https://api.github.com/users/macks22/orgs", "repos_url": "https://api.github.com/users/macks22/repos", "events_url": "https://api.github.com/users/macks22/events{/privacy}", "received_events_url": "https://api.github.com/users/macks22/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You can use the fields constraints param to specify default values: DateTimeField(constraints=(SQL('default now'),))\n\nI do not plan on supporting triggers at this time.\n", "Got it; thanks for the constraints tip.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/786/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "pc10201", "id": 8475089, "node_id": "MDQ6VXNlcjg0NzUwODk=", "avatar_url": "https://avatars.githubusercontent.com/u/8475089?v=4", "gravatar_id": "", "url": "https://api.github.com/users/pc10201", "html_url": "https://github.com/pc10201", "followers_url": "https://api.github.com/users/pc10201/followers", "following_url": "https://api.github.com/users/pc10201/following{/other_user}", "gists_url": "https://api.github.com/users/pc10201/gists{/gist_id}", "starred_url": "https://api.github.com/users/pc10201/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/pc10201/subscriptions", "organizations_url": "https://api.github.com/users/pc10201/orgs", "repos_url": "https://api.github.com/users/pc10201/repos", "events_url": "https://api.github.com/users/pc10201/events{/privacy}", "received_events_url": "https://api.github.com/users/pc10201/received_events", "type": "User", "site_admin": false }

[]

[]

[ "http://docs.peewee-orm.com/en/latest/peewee/api.html#InsertQuery.on_conflict\n\nPlease search the docs before posting tickets like this. Github tickets are more appropriate for **bugs**. If you have a question, post to stackoverflow, the mailing list, or the IRC channel. All of these are described in the documentation: http://docs.peewee-orm.com/en/latest/peewee/contributing.html\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/785/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "pc10201", "id": 8475089, "node_id": "MDQ6VXNlcjg0NzUwODk=", "avatar_url": "https://avatars.githubusercontent.com/u/8475089?v=4", "gravatar_id": "", "url": "https://api.github.com/users/pc10201", "html_url": "https://github.com/pc10201", "followers_url": "https://api.github.com/users/pc10201/followers", "following_url": "https://api.github.com/users/pc10201/following{/other_user}", "gists_url": "https://api.github.com/users/pc10201/gists{/gist_id}", "starred_url": "https://api.github.com/users/pc10201/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/pc10201/subscriptions", "organizations_url": "https://api.github.com/users/pc10201/orgs", "repos_url": "https://api.github.com/users/pc10201/repos", "events_url": "https://api.github.com/users/pc10201/events{/privacy}", "received_events_url": "https://api.github.com/users/pc10201/received_events", "type": "User", "site_admin": false }

[]

[]

[ "The idea is that it is bad practice to accept passwords on the command-line, since they get written to your bash history, etc.\n", "So you specify `-P`, then when prompted you enter your password in.\n", "Interesting that you chose not to accept the passwords on the command line which is an explicit choice that the end user makes, however the pwiz tool itself _echos the password into the console or is piped into the file you store the models in without explicitly telling the user that their password is going to get written_. This has led me to accidentally commit a password into a repo. You should be able to specify os.env variables to retrieve the passwords in the generated files, or else find some other paradigm to set the database on the BaseModel.", "> This has led me to accidentally commit a password into a repo.\r\n\r\nWhoops, I can add a note to the docs that the password will be outputted as a connection parameter.\r\n\r\nAs far as using an environment variable, just modify the generated code to do what you want it to.", "I am using peewee in a containerized project. There is a console program that manages the project, and I just want to run pwiz in a container where my python backend lives to get an updated schema from the database when developers change it.\r\n\r\nAll the database credentials are env variables inside the container, and this is a standard practice. We are working with dev environment with pwiz, as expected.\r\n\r\nInstead of simply running something like\r\n\r\n`docker exec <CONTAINER_ID> 'python -m pwiz -H $DBHOST -p $DBPORT -u $DBUSER -p $DBPASSWORD -e postgresql $DBNAME'`\r\n\r\nI have to ask for password or have to write an expect script just because there is no way for pwiz to accept a password non-interactively.\r\n\r\nThis is to show that enforcing users to enter password from the terminal is a flawed practice, and this is an issue we are facing with this tool, otherwise excellent. Maybe you reconsider your policy, thanks!\r\n", "`pwiz` is not typically something intended for use outside of interactive sessions, since it is really just a generator for skeleton code. You could, for your purposes, probably just look at the code inside pwiz's main and write your own short script." ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/784/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "bakemecookies", "id": 2167033, "node_id": "MDQ6VXNlcjIxNjcwMzM=", "avatar_url": "https://avatars.githubusercontent.com/u/2167033?v=4", "gravatar_id": "", "url": "https://api.github.com/users/bakemecookies", "html_url": "https://github.com/bakemecookies", "followers_url": "https://api.github.com/users/bakemecookies/followers", "following_url": "https://api.github.com/users/bakemecookies/following{/other_user}", "gists_url": "https://api.github.com/users/bakemecookies/gists{/gist_id}", "starred_url": "https://api.github.com/users/bakemecookies/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/bakemecookies/subscriptions", "organizations_url": "https://api.github.com/users/bakemecookies/orgs", "repos_url": "https://api.github.com/users/bakemecookies/repos", "events_url": "https://api.github.com/users/bakemecookies/events{/privacy}", "received_events_url": "https://api.github.com/users/bakemecookies/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Unfortunately, I don't think I will be going back to the \"sql error handler\" API. I think the implementation in the `RetryOperationalError` mixin is an appropriate way to try to handle re-connect/re-try logic. Maybe you can make some changes to your pool implementation, particularly overriding the `_is_closed()` method to perform some sort of test to check that the conn is really active. Take a look at the pool module's code, particularly the `PooledMySQLDatabase._is_closed()` method...I bet you can find something that'll work for you.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/783/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "xr09", "id": 1380243, "node_id": "MDQ6VXNlcjEzODAyNDM=", "avatar_url": "https://avatars.githubusercontent.com/u/1380243?v=4", "gravatar_id": "", "url": "https://api.github.com/users/xr09", "html_url": "https://github.com/xr09", "followers_url": "https://api.github.com/users/xr09/followers", "following_url": "https://api.github.com/users/xr09/following{/other_user}", "gists_url": "https://api.github.com/users/xr09/gists{/gist_id}", "starred_url": "https://api.github.com/users/xr09/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/xr09/subscriptions", "organizations_url": "https://api.github.com/users/xr09/orgs", "repos_url": "https://api.github.com/users/xr09/repos", "events_url": "https://api.github.com/users/xr09/events{/privacy}", "received_events_url": "https://api.github.com/users/xr09/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Thnx\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/782/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/782", "html_url": "https://github.com/coleifer/peewee/pull/782", "diff_url": "https://github.com/coleifer/peewee/pull/782.diff", "patch_url": "https://github.com/coleifer/peewee/pull/782.patch", "merged_at": "2015-12-04T02:53:09" }

{ "login": "mfussenegger", "id": 38700, "node_id": "MDQ6VXNlcjM4NzAw", "avatar_url": "https://avatars.githubusercontent.com/u/38700?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mfussenegger", "html_url": "https://github.com/mfussenegger", "followers_url": "https://api.github.com/users/mfussenegger/followers", "following_url": "https://api.github.com/users/mfussenegger/following{/other_user}", "gists_url": "https://api.github.com/users/mfussenegger/gists{/gist_id}", "starred_url": "https://api.github.com/users/mfussenegger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mfussenegger/subscriptions", "organizations_url": "https://api.github.com/users/mfussenegger/orgs", "repos_url": "https://api.github.com/users/mfussenegger/repos", "events_url": "https://api.github.com/users/mfussenegger/events{/privacy}", "received_events_url": "https://api.github.com/users/mfussenegger/received_events", "type": "User", "site_admin": false }

[]

[]

[]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/781/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "foxx", "id": 651797, "node_id": "MDQ6VXNlcjY1MTc5Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/651797?v=4", "gravatar_id": "", "url": "https://api.github.com/users/foxx", "html_url": "https://github.com/foxx", "followers_url": "https://api.github.com/users/foxx/followers", "following_url": "https://api.github.com/users/foxx/following{/other_user}", "gists_url": "https://api.github.com/users/foxx/gists{/gist_id}", "starred_url": "https://api.github.com/users/foxx/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/foxx/subscriptions", "organizations_url": "https://api.github.com/users/foxx/orgs", "repos_url": "https://api.github.com/users/foxx/repos", "events_url": "https://api.github.com/users/foxx/events{/privacy}", "received_events_url": "https://api.github.com/users/foxx/received_events", "type": "User", "site_admin": false }

[]

[]

[ "LGTM, you're too fast!!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/780/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "amorphic", "id": 1538480, "node_id": "MDQ6VXNlcjE1Mzg0ODA=", "avatar_url": "https://avatars.githubusercontent.com/u/1538480?v=4", "gravatar_id": "", "url": "https://api.github.com/users/amorphic", "html_url": "https://github.com/amorphic", "followers_url": "https://api.github.com/users/amorphic/followers", "following_url": "https://api.github.com/users/amorphic/following{/other_user}", "gists_url": "https://api.github.com/users/amorphic/gists{/gist_id}", "starred_url": "https://api.github.com/users/amorphic/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/amorphic/subscriptions", "organizations_url": "https://api.github.com/users/amorphic/orgs", "repos_url": "https://api.github.com/users/amorphic/repos", "events_url": "https://api.github.com/users/amorphic/events{/privacy}", "received_events_url": "https://api.github.com/users/amorphic/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Apologies... this had absolutely nothing to do with peewee.\n", "What was the actual problem? I have a similar error.\n", "I honestly can't remember but I was doing something stupid which was unrelated to peewee.\n\nThe module I was working on is [here](https://github.com/re-3D/opengb/blob/master/opengb/database.py) in case it's useful.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/779/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "ccario83", "id": 3374143, "node_id": "MDQ6VXNlcjMzNzQxNDM=", "avatar_url": "https://avatars.githubusercontent.com/u/3374143?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ccario83", "html_url": "https://github.com/ccario83", "followers_url": "https://api.github.com/users/ccario83/followers", "following_url": "https://api.github.com/users/ccario83/following{/other_user}", "gists_url": "https://api.github.com/users/ccario83/gists{/gist_id}", "starred_url": "https://api.github.com/users/ccario83/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/ccario83/subscriptions", "organizations_url": "https://api.github.com/users/ccario83/orgs", "repos_url": "https://api.github.com/users/ccario83/repos", "events_url": "https://api.github.com/users/ccario83/events{/privacy}", "received_events_url": "https://api.github.com/users/ccario83/received_events", "type": "User", "site_admin": false }

[]

[]

[ "> But wouldn't this sort of boilerplate code be better included in peewee itself?\n\nNot in my opinion, no. There are far too many things involved for me to feel comfortable implementing a canonical / library solution to the problem of deadlocks.\n", "@lucidv01d Did you implement some generic(ish) retry logic or how did you end up tackling it?" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/778/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "alexanderad", "id": 888394, "node_id": "MDQ6VXNlcjg4ODM5NA==", "avatar_url": "https://avatars.githubusercontent.com/u/888394?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alexanderad", "html_url": "https://github.com/alexanderad", "followers_url": "https://api.github.com/users/alexanderad/followers", "following_url": "https://api.github.com/users/alexanderad/following{/other_user}", "gists_url": "https://api.github.com/users/alexanderad/gists{/gist_id}", "starred_url": "https://api.github.com/users/alexanderad/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/alexanderad/subscriptions", "organizations_url": "https://api.github.com/users/alexanderad/orgs", "repos_url": "https://api.github.com/users/alexanderad/repos", "events_url": "https://api.github.com/users/alexanderad/events{/privacy}", "received_events_url": "https://api.github.com/users/alexanderad/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Ah, `naive()` can help. Sorry.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/777/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "alexanderad", "id": 888394, "node_id": "MDQ6VXNlcjg4ODM5NA==", "avatar_url": "https://avatars.githubusercontent.com/u/888394?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alexanderad", "html_url": "https://github.com/alexanderad", "followers_url": "https://api.github.com/users/alexanderad/followers", "following_url": "https://api.github.com/users/alexanderad/following{/other_user}", "gists_url": "https://api.github.com/users/alexanderad/gists{/gist_id}", "starred_url": "https://api.github.com/users/alexanderad/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/alexanderad/subscriptions", "organizations_url": "https://api.github.com/users/alexanderad/orgs", "repos_url": "https://api.github.com/users/alexanderad/repos", "events_url": "https://api.github.com/users/alexanderad/events{/privacy}", "received_events_url": "https://api.github.com/users/alexanderad/received_events", "type": "User", "site_admin": false }

[]

[]

[ "It seems pretty straightforward to me:\n\n``` python\nIn [15]: with assert_query_count(1):\n ....: q = User.select(User, Group).join(Group)\n ....: for user in q:\n ....: print user.username, user.group.name\n ....: \n... prints users and groups ...\n```\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/776/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "hedger", "id": 277532, "node_id": "MDQ6VXNlcjI3NzUzMg==", "avatar_url": "https://avatars.githubusercontent.com/u/277532?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hedger", "html_url": "https://github.com/hedger", "followers_url": "https://api.github.com/users/hedger/followers", "following_url": "https://api.github.com/users/hedger/following{/other_user}", "gists_url": "https://api.github.com/users/hedger/gists{/gist_id}", "starred_url": "https://api.github.com/users/hedger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/hedger/subscriptions", "organizations_url": "https://api.github.com/users/hedger/orgs", "repos_url": "https://api.github.com/users/hedger/repos", "events_url": "https://api.github.com/users/hedger/events{/privacy}", "received_events_url": "https://api.github.com/users/hedger/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Yay, all checks have passed.\n", "I've pushed a branch that has some of your changes integrated, 89cd2c4f74a526591f3804d04b2252c489be5432\n", "That's great, thanks!\n", "Pushed 2f028c32225c5aad9b97f0f451b9d238c3846e0d to the `multiprefetch` branch. Things aren't working quite right in the test-case. I think the identity map needs to differentiate things by field, but haven't dug into it too much yet.\n", "Got things fixed and have merged into master.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/775/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/775", "html_url": "https://github.com/coleifer/peewee/pull/775", "diff_url": "https://github.com/coleifer/peewee/pull/775.diff", "patch_url": "https://github.com/coleifer/peewee/pull/775.patch", "merged_at": null }

{ "login": "hedger", "id": 277532, "node_id": "MDQ6VXNlcjI3NzUzMg==", "avatar_url": "https://avatars.githubusercontent.com/u/277532?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hedger", "html_url": "https://github.com/hedger", "followers_url": "https://api.github.com/users/hedger/followers", "following_url": "https://api.github.com/users/hedger/following{/other_user}", "gists_url": "https://api.github.com/users/hedger/gists{/gist_id}", "starred_url": "https://api.github.com/users/hedger/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/hedger/subscriptions", "organizations_url": "https://api.github.com/users/hedger/orgs", "repos_url": "https://api.github.com/users/hedger/repos", "events_url": "https://api.github.com/users/hedger/events{/privacy}", "received_events_url": "https://api.github.com/users/hedger/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Tried to fix that, https://github.com/coleifer/peewee/pull/775.\nTested on some backrefs in my project. Still needs more testing.\n", "Fix merged.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/774/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "perplexedfingers", "id": 14918698, "node_id": "MDQ6VXNlcjE0OTE4Njk4", "avatar_url": "https://avatars.githubusercontent.com/u/14918698?v=4", "gravatar_id": "", "url": "https://api.github.com/users/perplexedfingers", "html_url": "https://github.com/perplexedfingers", "followers_url": "https://api.github.com/users/perplexedfingers/followers", "following_url": "https://api.github.com/users/perplexedfingers/following{/other_user}", "gists_url": "https://api.github.com/users/perplexedfingers/gists{/gist_id}", "starred_url": "https://api.github.com/users/perplexedfingers/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/perplexedfingers/subscriptions", "organizations_url": "https://api.github.com/users/perplexedfingers/orgs", "repos_url": "https://api.github.com/users/perplexedfingers/repos", "events_url": "https://api.github.com/users/perplexedfingers/events{/privacy}", "received_events_url": "https://api.github.com/users/perplexedfingers/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Sure thanks!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/773/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "mekanix", "id": 610855, "node_id": "MDQ6VXNlcjYxMDg1NQ==", "avatar_url": "https://avatars.githubusercontent.com/u/610855?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mekanix", "html_url": "https://github.com/mekanix", "followers_url": "https://api.github.com/users/mekanix/followers", "following_url": "https://api.github.com/users/mekanix/following{/other_user}", "gists_url": "https://api.github.com/users/mekanix/gists{/gist_id}", "starred_url": "https://api.github.com/users/mekanix/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mekanix/subscriptions", "organizations_url": "https://api.github.com/users/mekanix/orgs", "repos_url": "https://api.github.com/users/mekanix/repos", "events_url": "https://api.github.com/users/mekanix/events{/privacy}", "received_events_url": "https://api.github.com/users/mekanix/received_events", "type": "User", "site_admin": false }

[]

[]

[ "@mekanix -- https://github.com/cam-stitt/arnold is one such project I'm aware of. I don't have plans for adding anything more than what is already present in the `playhouse.migrate` module. Perhaps you can bring this up on the mailing list? https://groups.google.com/group/peewee-orm\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/772/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "rlander", "id": 368935, "node_id": "MDQ6VXNlcjM2ODkzNQ==", "avatar_url": "https://avatars.githubusercontent.com/u/368935?v=4", "gravatar_id": "", "url": "https://api.github.com/users/rlander", "html_url": "https://github.com/rlander", "followers_url": "https://api.github.com/users/rlander/followers", "following_url": "https://api.github.com/users/rlander/following{/other_user}", "gists_url": "https://api.github.com/users/rlander/gists{/gist_id}", "starred_url": "https://api.github.com/users/rlander/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/rlander/subscriptions", "organizations_url": "https://api.github.com/users/rlander/orgs", "repos_url": "https://api.github.com/users/rlander/repos", "events_url": "https://api.github.com/users/rlander/events{/privacy}", "received_events_url": "https://api.github.com/users/rlander/received_events", "type": "User", "site_admin": false }

[]

[]

[ "```\nPython version: 2.7.9\nSqlite version: 3.9.2\n\npeewee==2.7.3\napsw==3.9.2-r1\n\n\n```\n", "You need to use the `DateTimeField` object from the `apsw_ext` module.\n\n``` python\nfrom playhouse.apsw_ext import DateTimeField\n```\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/771/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "JornPet", "id": 9035151, "node_id": "MDQ6VXNlcjkwMzUxNTE=", "avatar_url": "https://avatars.githubusercontent.com/u/9035151?v=4", "gravatar_id": "", "url": "https://api.github.com/users/JornPet", "html_url": "https://github.com/JornPet", "followers_url": "https://api.github.com/users/JornPet/followers", "following_url": "https://api.github.com/users/JornPet/following{/other_user}", "gists_url": "https://api.github.com/users/JornPet/gists{/gist_id}", "starred_url": "https://api.github.com/users/JornPet/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/JornPet/subscriptions", "organizations_url": "https://api.github.com/users/JornPet/orgs", "repos_url": "https://api.github.com/users/JornPet/repos", "events_url": "https://api.github.com/users/JornPet/events{/privacy}", "received_events_url": "https://api.github.com/users/JornPet/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I'm not sure, maybe something holding a lock on the table? Are any other processes connected to the db?\n", "Closing due to lack of feedback.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/770/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "martinburch", "id": 2335284, "node_id": "MDQ6VXNlcjIzMzUyODQ=", "avatar_url": "https://avatars.githubusercontent.com/u/2335284?v=4", "gravatar_id": "", "url": "https://api.github.com/users/martinburch", "html_url": "https://github.com/martinburch", "followers_url": "https://api.github.com/users/martinburch/followers", "following_url": "https://api.github.com/users/martinburch/following{/other_user}", "gists_url": "https://api.github.com/users/martinburch/gists{/gist_id}", "starred_url": "https://api.github.com/users/martinburch/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/martinburch/subscriptions", "organizations_url": "https://api.github.com/users/martinburch/orgs", "repos_url": "https://api.github.com/users/martinburch/repos", "events_url": "https://api.github.com/users/martinburch/events{/privacy}", "received_events_url": "https://api.github.com/users/martinburch/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Peewee leans on pythons parser to generate expressions. Because of pythons operator precedence you need parentheses around most comparisons. No magic here, just python.\n", "And yet I can do this without needed extra parens:\n\n```\n.where(\n (modelObject.field_name << ['a','b','c'] |\n modelObject.field_name % \"%word%\")\n )\n```\n\nSeems inconsistent. I think this needs to be explained in the docs. At least \"you should use parentheses around all comparisons, just to be safe.\" But a list of situations would be ideal.\n", "What do you suggest? Best practice is just put parentheses around binary expressions, then you won't have any issues.\n", "You can created the equivalent expressions by calling the appropriate doubles underscore method on the field object. You can also instantiate expression objects by hand.\n", "The use of parentheses in your python code won't affect the SQL being generated, improves readability for complex, mutli-clause queries (as I typically assume you line-break each expression), and leaves no room for error.\n\nhttp://docs.peewee-orm.com/en/latest/peewee/querying.html?highlight=precedence#expressions\n", "Sorry, I think you've already got a note in the docs to this effect.\n\n> Note that the actual comparisons are wrapped in parentheses. Python’s operator precedence necessitates that comparisons be wrapped in parentheses.\n\nThanks for your additional thoughts on this solidly closed issue.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/769/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "klen", "id": 90699, "node_id": "MDQ6VXNlcjkwNjk5", "avatar_url": "https://avatars.githubusercontent.com/u/90699?v=4", "gravatar_id": "", "url": "https://api.github.com/users/klen", "html_url": "https://github.com/klen", "followers_url": "https://api.github.com/users/klen/followers", "following_url": "https://api.github.com/users/klen/following{/other_user}", "gists_url": "https://api.github.com/users/klen/gists{/gist_id}", "starred_url": "https://api.github.com/users/klen/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/klen/subscriptions", "organizations_url": "https://api.github.com/users/klen/orgs", "repos_url": "https://api.github.com/users/klen/repos", "events_url": "https://api.github.com/users/klen/events{/privacy}", "received_events_url": "https://api.github.com/users/klen/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Thank you!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/768/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/768", "html_url": "https://github.com/coleifer/peewee/pull/768", "diff_url": "https://github.com/coleifer/peewee/pull/768.diff", "patch_url": "https://github.com/coleifer/peewee/pull/768.patch", "merged_at": "2015-12-01T15:08:48" }

{ "login": "felixonmars", "id": 1006477, "node_id": "MDQ6VXNlcjEwMDY0Nzc=", "avatar_url": "https://avatars.githubusercontent.com/u/1006477?v=4", "gravatar_id": "", "url": "https://api.github.com/users/felixonmars", "html_url": "https://github.com/felixonmars", "followers_url": "https://api.github.com/users/felixonmars/followers", "following_url": "https://api.github.com/users/felixonmars/following{/other_user}", "gists_url": "https://api.github.com/users/felixonmars/gists{/gist_id}", "starred_url": "https://api.github.com/users/felixonmars/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/felixonmars/subscriptions", "organizations_url": "https://api.github.com/users/felixonmars/orgs", "repos_url": "https://api.github.com/users/felixonmars/repos", "events_url": "https://api.github.com/users/felixonmars/events{/privacy}", "received_events_url": "https://api.github.com/users/felixonmars/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Looks like you're missing the JSON1 extension. I'll add a conditional.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/767/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Thanks :)\n", "Same code, with updated peewee:\n\n```\nTraceback (most recent call last):\n File \"peewee_password_field.py\", line 19, in <module>\n assert user.password.check_password('password')\n File \"/home/ybon/.virtualenvs/p3/local/lib/python3.4/site-packages/playhouse/fields.py\", line 47, in check_password\n return hashpw(password, self) == self\n File \"/home/ybon/.virtualenvs/p3/local/lib/python3.4/site-packages/bcrypt/__init__.py\", line 66, in hashpw\n raise ValueError(\"Invalid salt\")\nValueError: Invalid salt\n```\n\nMissing unittests in Postgres? :)\n", "(I can't reopen the issue, dunno why…)\n", "I'm not sure what's going on exactly. Having trouble getting `bcrypt` to build in a python3 virtualenv but testing locally with python2+sqlite is working fine. What is the salt of the particular password you're having trouble with, and what is it's type?\n", "Thanks for looking into this @coleifer :)\n\nI'm just running the snippet from https://github.com/coleifer/peewee/issues/766#issue-118223951 (I'm on IRC if it helps debugging).\n", "This diff fixes my usecase:\n\n``` diff\ndiff --git a/playhouse/fields.py b/playhouse/fields.py\nindex fcebae7..bfa2d90 100644\n--- a/playhouse/fields.py\n+++ b/playhouse/fields.py\n@@ -47,7 +47,7 @@ if hashpw and gensalt:\n return hashpw(password, self) == self\n\n\n- class PasswordField(TextField):\n+ class PasswordField(BlobField):\n def __init__(self, iterations=12, *args, **kwargs):\n if None in (hashpw, gensalt):\n raise ValueError('Missing library required for PasswordField: bcrypt')\n```\n\nBasically, the issue seems to be that we are sending bytes to a Text field in Postgres.\nNot sure though why tests passe without this change (they also pass with the fix)…\n", "Ah, now I know why tests passe even without the fix: they are never ran with postgres…\nEven if I pass `--engine postgres`, simply because the Model used in the tests only use \"in memory database\": https://github.com/coleifer/peewee/blob/master/playhouse/tests/test_extra_fields.py#L22\n\nIs there a reason for that?\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/766/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "pawl", "id": 992533, "node_id": "MDQ6VXNlcjk5MjUzMw==", "avatar_url": "https://avatars.githubusercontent.com/u/992533?v=4", "gravatar_id": "", "url": "https://api.github.com/users/pawl", "html_url": "https://github.com/pawl", "followers_url": "https://api.github.com/users/pawl/followers", "following_url": "https://api.github.com/users/pawl/following{/other_user}", "gists_url": "https://api.github.com/users/pawl/gists{/gist_id}", "starred_url": "https://api.github.com/users/pawl/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/pawl/subscriptions", "organizations_url": "https://api.github.com/users/pawl/orgs", "repos_url": "https://api.github.com/users/pawl/repos", "events_url": "https://api.github.com/users/pawl/events{/privacy}", "received_events_url": "https://api.github.com/users/pawl/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I'd look into updating the code to use `.sorted_fields`. there is also `.sorted_field_names`. This is a performance improvement, as that code had potential to be called a _lot_.\n", "@rudyryk has a good suggestion here for a temporary solution: https://github.com/flask-admin/flask-admin/issues/1117\n\n```\n for n, f in zip(model._meta.sorted_field_names,\n model._meta.sorted_fields):\n```\n", "Folks, I'm still getting this error.\r\n\r\nTraceback:\r\n```\r\n File \"app.py\", line 30, in <module>\r\n admin.add_view(ModelView(database.Comic))\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/contrib/peewee/view.py\", line 144, in __init__\r\n menu_icon_value=menu_icon_value)\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/model/base.py\", line 584, in __init__\r\n self._refresh_cache()\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/model/base.py\", line 653, in _refresh_cache\r\n self._list_columns = self.get_list_columns()\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/model/base.py\", line 729, in get_list_columns\r\n columns = self.scaffold_list_columns()\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/contrib/peewee/view.py\", line 163, in scaffold_list_columns\r\n for n, f in self._get_model_fields():\r\n File \"/usr/local/lib/python3.5/dist-packages/flask_admin/contrib/peewee/view.py\", line 152, in _get_model_fields\r\n return model._meta.get_sorted_fields()\r\nAttributeError: 'ModelOptions' object has no attribute 'get_sorted_fields'\r\n```\r\n\r\nVersions:\r\n```\r\nFlask==0.12.2\r\nFlask-Admin==1.1.0\r\npeewee==2.10.1\r\nwtf-peewee==0.2.6\r\n```\r\n\r\nWhat is going on?\r\n\r\n\r\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/765/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "olivetree123", "id": 5275647, "node_id": "MDQ6VXNlcjUyNzU2NDc=", "avatar_url": "https://avatars.githubusercontent.com/u/5275647?v=4", "gravatar_id": "", "url": "https://api.github.com/users/olivetree123", "html_url": "https://github.com/olivetree123", "followers_url": "https://api.github.com/users/olivetree123/followers", "following_url": "https://api.github.com/users/olivetree123/following{/other_user}", "gists_url": "https://api.github.com/users/olivetree123/gists{/gist_id}", "starred_url": "https://api.github.com/users/olivetree123/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/olivetree123/subscriptions", "organizations_url": "https://api.github.com/users/olivetree123/orgs", "repos_url": "https://api.github.com/users/olivetree123/repos", "events_url": "https://api.github.com/users/olivetree123/events{/privacy}", "received_events_url": "https://api.github.com/users/olivetree123/received_events", "type": "User", "site_admin": false }

[]

[]

[ "No, that has been that way since version 0.1.0. Are you serious? Do you think I'd actually change it?\n", "Also, cluttering up the namespace is grosser to me.\n", "Also bite me.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/764/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "mindv0rtex", "id": 10409817, "node_id": "MDQ6VXNlcjEwNDA5ODE3", "avatar_url": "https://avatars.githubusercontent.com/u/10409817?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mindv0rtex", "html_url": "https://github.com/mindv0rtex", "followers_url": "https://api.github.com/users/mindv0rtex/followers", "following_url": "https://api.github.com/users/mindv0rtex/following{/other_user}", "gists_url": "https://api.github.com/users/mindv0rtex/gists{/gist_id}", "starred_url": "https://api.github.com/users/mindv0rtex/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mindv0rtex/subscriptions", "organizations_url": "https://api.github.com/users/mindv0rtex/orgs", "repos_url": "https://api.github.com/users/mindv0rtex/repos", "events_url": "https://api.github.com/users/mindv0rtex/events{/privacy}", "received_events_url": "https://api.github.com/users/mindv0rtex/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Hmm, well this isn't exactly a solution, but I've released a new versoin 2.7.2 that does not build SQLite extension by default.\n\nIf you want the sqlite extension it should be built using `python setup.py build_sqlite_ext`.\n\nI will look into checking the python version and renaming the entry-point per the stackoverflow answer. (`PyInit__sqlite_ext`).\n\nFuck python 3 and it's stupid ass changes.\n", "FWIW, I confirm that I could successfully perform the `python setup.py build_sqlite_ext` under WinPython 2, so this indeed looks like a Python 3 issue.\n", "Some issues with the `_sqlite_ext` extension:\n1. On Python 3, the extension module fails to link. The module's init function is actually named `PyInit__sqlite_ext`, not [`init_sqlite_ext`](https://github.com/coleifer/peewee/blob/master/playhouse/_sqlite_ext.pyx#L289)\n2. On Python 3 for Windows, the [`sqlite3_sqliteext_init`](https://github.com/coleifer/peewee/blob/master/playhouse/_sqlite_ext.pyx#L283) function is not exported from the module. I think this is a Cython bug.\n3. On Windows, the [`load_extension`](https://github.com/coleifer/peewee/blob/master/playhouse/sqlite_ext.py#L711) function fails to load the `_sqlite_ext` module because the file extension of the Python extension module is `.pyd`, not `.dll`. Also, on Python 3.5, the name of the Python extension is `_sqlite_ext.cp35-win_amd64.pyd` or `_sqlite_ext.cp35-win32.pyd`, not `_sqlite_ext.pyd`\n", "Hello, encountered this problem again (sorry for the duplicate issue). Would there be a fix, other than to not use peewee with Py3.X on windows?\n", "@shirish93 can you please paste the traceback?\n", "I solved the issue (probably more #831) with my python 3.5 setup by insalling VS 2015 with the instructions of this thread:\n\nhttp://stackoverflow.com/questions/33323172/vcvarsall-bat-needed-for-python-to-compile-missing-from-visual-studio-2015-v-1\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/763/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "skatenerd", "id": 880217, "node_id": "MDQ6VXNlcjg4MDIxNw==", "avatar_url": "https://avatars.githubusercontent.com/u/880217?v=4", "gravatar_id": "", "url": "https://api.github.com/users/skatenerd", "html_url": "https://github.com/skatenerd", "followers_url": "https://api.github.com/users/skatenerd/followers", "following_url": "https://api.github.com/users/skatenerd/following{/other_user}", "gists_url": "https://api.github.com/users/skatenerd/gists{/gist_id}", "starred_url": "https://api.github.com/users/skatenerd/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/skatenerd/subscriptions", "organizations_url": "https://api.github.com/users/skatenerd/orgs", "repos_url": "https://api.github.com/users/skatenerd/repos", "events_url": "https://api.github.com/users/skatenerd/events{/privacy}", "received_events_url": "https://api.github.com/users/skatenerd/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Just use `%%` instead of `%`\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/762/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "openmartin", "id": 2059728, "node_id": "MDQ6VXNlcjIwNTk3Mjg=", "avatar_url": "https://avatars.githubusercontent.com/u/2059728?v=4", "gravatar_id": "", "url": "https://api.github.com/users/openmartin", "html_url": "https://github.com/openmartin", "followers_url": "https://api.github.com/users/openmartin/followers", "following_url": "https://api.github.com/users/openmartin/following{/other_user}", "gists_url": "https://api.github.com/users/openmartin/gists{/gist_id}", "starred_url": "https://api.github.com/users/openmartin/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/openmartin/subscriptions", "organizations_url": "https://api.github.com/users/openmartin/orgs", "repos_url": "https://api.github.com/users/openmartin/repos", "events_url": "https://api.github.com/users/openmartin/events{/privacy}", "received_events_url": "https://api.github.com/users/openmartin/received_events", "type": "User", "site_admin": false }

[]

[]

[ "on Ubuntu 14.04.2 LTS\npython 2.7.6\n", "I confirm we got the same issue this morning on our CI.\nOur trace is:\n\n`````` python\nDownloading/unpacking peewee>=2.6.4 (from ...->-r /home/u877/src/repo/app-api/requirements.txt (line 7))\n Running setup.py (path:/home/u877/python-env/build/peewee/setup.py) egg_info for package peewee\n Traceback (most recent call last):\n File \"<string>\", line 17, in <module>\n File \"/home/u877/python-env/build/peewee/setup.py\", line 42, in <module>\n scripts = ['pwiz.py'],\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/core.py\", line 138, in setup\n ok = dist.parse_command_line()\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 250, in parse_command_line\n result = _Distribution.parse_command_line(self)\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/dist.py\", line 467, in parse_command_line\n args = self._parse_command_opts(parser, args)\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 534, in _parse_command_opts\n nargs = _Distribution._parse_command_opts(self, parser, args)\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/dist.py\", line 523, in _parse_command_opts\n cmd_class = self.get_command_class(command)\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 359, in get_command_class\n if command in self.cmdclass:\n TypeError: argument of type 'NoneType' is not iterable\n Complete output from command python setup.py egg_info:\n Traceback (most recent call last):\n\n File \"<string>\", line 17, in <module>\n\n File \"/home/u877/python-env/build/peewee/setup.py\", line 42, in <module>\n\n scripts = ['pwiz.py'],\n\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/core.py\", line 138, in setup\n\n ok = dist.parse_command_line()\n\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 250, in parse_command_line\n\n result = _Distribution.parse_command_line(self)\n\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/dist.py\", line 467, in parse_command_line\n\n args = self._parse_command_opts(parser, args)\n\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 534, in _parse_command_opts\n\n nargs = _Distribution._parse_command_opts(self, parser, args)\n\n File \"/usr/local/python/cpython-2.7/lib/python2.7/distutils/dist.py\", line 523, in _parse_command_opts\n\n cmd_class = self.get_command_class(command)\n\n File \"/home/u877/python-env/lib/python2.7/site-packages/setuptools/dist.py\", line 359, in get_command_class\n\n if command in self.cmdclass:\n\nTypeError: argument of type 'NoneType' is not iterable\n\n----------------------------------------\nCleaning up...\nCommand python setup.py egg_info failed with error code 1 in /home/u877/python-env/build/peewee\nStoring debug log for failure in /home/u877/.pip/pip.log\n[exit status: 1]```\n\n``````\n", ":+1: Python 3.4.2 https://travis-ci.org/klen/muffin-admin/builds/92235523\n\nLooks like setup with peewee 2.7.0 is broken.\n", "Same error occurs on Mac OS X 10.11, Python 3.5\n", "This should be fixed with the 2.7.1 release. Tested with and without cython on py 2.7.9.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/761/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "olivetree123", "id": 5275647, "node_id": "MDQ6VXNlcjUyNzU2NDc=", "avatar_url": "https://avatars.githubusercontent.com/u/5275647?v=4", "gravatar_id": "", "url": "https://api.github.com/users/olivetree123", "html_url": "https://github.com/olivetree123", "followers_url": "https://api.github.com/users/olivetree123/followers", "following_url": "https://api.github.com/users/olivetree123/following{/other_user}", "gists_url": "https://api.github.com/users/olivetree123/gists{/gist_id}", "starred_url": "https://api.github.com/users/olivetree123/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/olivetree123/subscriptions", "organizations_url": "https://api.github.com/users/olivetree123/orgs", "repos_url": "https://api.github.com/users/olivetree123/repos", "events_url": "https://api.github.com/users/olivetree123/events{/privacy}", "received_events_url": "https://api.github.com/users/olivetree123/received_events", "type": "User", "site_admin": false }

[]

[]

[ "First off this is completely inappropriate for a GitHub issue. Questions go on stack overflow, irc or the mailing list.\n\nSecond, to answer, write:\n\nGym.select().dicts().get()\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/760/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "skatenerd", "id": 880217, "node_id": "MDQ6VXNlcjg4MDIxNw==", "avatar_url": "https://avatars.githubusercontent.com/u/880217?v=4", "gravatar_id": "", "url": "https://api.github.com/users/skatenerd", "html_url": "https://github.com/skatenerd", "followers_url": "https://api.github.com/users/skatenerd/followers", "following_url": "https://api.github.com/users/skatenerd/following{/other_user}", "gists_url": "https://api.github.com/users/skatenerd/gists{/gist_id}", "starred_url": "https://api.github.com/users/skatenerd/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/skatenerd/subscriptions", "organizations_url": "https://api.github.com/users/skatenerd/orgs", "repos_url": "https://api.github.com/users/skatenerd/repos", "events_url": "https://api.github.com/users/skatenerd/events{/privacy}", "received_events_url": "https://api.github.com/users/skatenerd/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Fixed with the 2.7.1 release. Cython is not required, but if you have it peewee will build some c extension.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/759/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "larsks", "id": 82622, "node_id": "MDQ6VXNlcjgyNjIy", "avatar_url": "https://avatars.githubusercontent.com/u/82622?v=4", "gravatar_id": "", "url": "https://api.github.com/users/larsks", "html_url": "https://github.com/larsks", "followers_url": "https://api.github.com/users/larsks/followers", "following_url": "https://api.github.com/users/larsks/following{/other_user}", "gists_url": "https://api.github.com/users/larsks/gists{/gist_id}", "starred_url": "https://api.github.com/users/larsks/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/larsks/subscriptions", "organizations_url": "https://api.github.com/users/larsks/orgs", "repos_url": "https://api.github.com/users/larsks/repos", "events_url": "https://api.github.com/users/larsks/events{/privacy}", "received_events_url": "https://api.github.com/users/larsks/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I would just create wrapper methods that do what you want, since by the time you're calling `python_value()` it's technically too late as the value has already been pulled from the database. As to `db_value()` doing the inverse, that's not really the intended purpose of the method.\n\nSince you're using Postgres, you can also register custom adapters: https://github.com/coleifer/peewee/blob/master/playhouse/postgres_ext.py#L106-L120\n", "@larsks FYI, I have the same need, and started working on https://github.com/yohanboniface/psycopg-postgis\nExample usage: https://github.com/etalab/ban/blob/master/ban/db/fields.py#L16\nI may wrap all this in a `peewee-postgis` module at some point, but for now it's experimental and I just need Point in my app. So work in progress… ;)\n", "Thanks for the pointer! I ultimately followed @coleifer suggestion and set up an adapter, which is pretty much straight out of the psycopg docs:\n\n```\nPoint = namedtuple('Point', ['x', 'y', 'srid'])\n\ndef adapt_point(point):\n x = adapt(point.x).getquoted()\n y = adapt(point.y).getquoted()\n srid = adapt(point.srid).getquoted()\n return AsIs(\"st_setsrid(st_makepoint(%s, %s), %s)\" % (x,y,srid))\n\nregister_adapter(Point, adapt_point)\n```\n\nAnd then in my PointField class:\n\n```\ndef db_value(self, value): \n if isinstance(value, tuple): \n return Point(value[0], value[1], self.geom_srid) \n else: \n return value \n```\n\nAnd this Just Works as far as storing data goes. It doesn't handle\nconverting the stored data back into Python tapes on load, because\nthis seemed a bit trickier, since it involves querying the database\nfor oid information.\n\nI also discovered the [shapely](https://pypi.python.org/pypi/Shapely) module, which has functions for\nhandling WKT/WKB data in Python, which would probably make this all\nmuch simpler (and less tied to psycopg-specific features).\n", "> It doesn't handle\n> converting the stored data back into Python tapes on load, because\n> this seemed a bit trickier, since it involves querying the database\n> for oid information.\n\nYep, that's what psycopg-postgis does. I need the conversion also on data fetching :)\n\n> I also discovered the shapely module, which has functions for\n> handling WKT/WKB data in Python\n\nAFAIK, it does not handle srid. :/\n", "@coleifer is there a way to override the fn called when performing a read?\r\n\r\nIE, instead of \r\n\r\n```sql\r\nselect foo, bar, baz from table\r\n```\r\n\r\nYou could specify a wrapper fn for a field, so you could so something like\r\n\r\n```sql\r\nselect foo, bar, ST_AsText(baz) from table\r\n```\r\n\r\nwhich would alleviate a lot of the trouble here (I too am trying to figure out how to read/write a simple `geometry(Point)` field with postgis.", "Have you tried implementing the `db_value` method such that it calls the function when passing the value into the db?\r\n\r\n```python\r\nclass GeometryField(WhateverField):\r\n def db_value(self, value):\r\n return fn.ST_AsText(value)\r\n```", "FYI, for those curious, the adapt code linked above was lost in recent code changes but I found it here in a 2.x release: https://github.com/coleifer/peewee/blob/761f9144a0e17381147a81658019cffe14c118ca/playhouse/postgres_ext.py#L106-L120", "As of commit 090dc61d08929bec863fc08da342284116807a42, the OP's original attempt should be working:\r\n\r\n```python\r\ndef db_value(self, value):\r\n return fn.st_setsrid(\r\n fn.st_makepoint(value[0],value[1],self.geom_srid))\r\n```\r\n\r\nAlthough if the value is a list, it may be necessary to wrap it in `AsIs()` additionally.", "First off, I want to say thanks to @coleifer for writing and maintaining such a great library.\r\n\r\nI think what @whalesalad is getting at is the same problem / hassle I'm running into: the extension (e.g. PostGIS) stores a different native type in the database, so when you want to use it outside the DB you need to serialize it to something like WKT. This has to be done at the time of query generation, as in their example:\r\n\r\n`select foo, bar, ST_AsText(baz) from table`\r\n\r\nI initially thought I could use fn.AsText (I'm trying this with Spatialite) in a `python_value` method, but as you pointed out this doesn't run until after the query and just ends up returning a useless function object.\r\n\r\nIf I understand right, the example given by @coleifer would wrap geometry fields with ST_AsText in INSERT and UPDATE statements in addition to SELECTS:\r\n\r\n class GeometryField(WhateverField):\r\n def db_value(self, value):\r\n return fn.ST_AsText(value)\r\n\r\nI think this would break INSERT and UPDATE statements attempting to store geometries in db-native format - it would write strings instead.\r\n\r\nAt the moment, I have a utility function that checks if a field name is 'geometry' and if so, wraps it in fn.AsText:\r\n\r\n def fix_geom_field(table, query_fields):\r\n fields = []\r\n for field_name in table._meta.sorted_field_names:\r\n if field_name in query_fields:\r\n field = getattr(table, field_name)\r\n if field_name == 'geometry':\r\n field = fn.AsText(field)\r\n fields.append(field)\r\n\r\nThis gets run on a list of field names that get converted into field instances associated with the table of interest, and wraps a column names geometry (this could also work with introspection into field types). The results of this query can then be used in ORM selects:\r\n\r\n Table.select(*fix_geom_field(Table, query_fields))", "Experienced developer learning PeeWee for the first time. Thank you so much, @coleifer.\r\n\r\nHas this updated in the last year? In @nbolten 's implementation, it does seem as though the map from field -> fn.AsText(field) ought to be able to be inserted into the GeometryField declaration class itself. ", "This appears to create the same outcome as @nbolten , and allows the containment to be on the field itself. It is unclear to me whether _returning is only those fields selected by the select() method, and how this will affect performance in more complex queries. Posting here, though.\r\n\r\n```\r\nclass PointField(pw.Field):\r\n \"\"\"Geometry field type Point, conforming to MySQL 8.0+, format is WKT: 'POINT(18 23)'\"\"\"\r\n\r\n # Field type in MySQL is Point with a capital P.\r\n field_type = 'Point'\r\n\r\n def python_value(self, value):\r\n return value\r\n\r\n def db_value(self, ewkt):\r\n \"\"\"Converts EWKT format to MySQL database requirements, e.g., SRID=3857;POINT(10,10)\"\"\"\r\n\r\n # Does the ewkt start with \"SRID=XXXX;...\".\r\n # If so, remove the value from the geometry and parse/extract the srid\r\n s = ewkt.split(\";\")\r\n if s[0].startswith(\"SRID=\"):\r\n geo = \";\".join(s[1:])\r\n srid = s[0].split(\"=\")[1]\r\n return pw.fn.ST_GeomFromText(geo, srid)\r\n else:\r\n return pw.fn.ST_GeomFromText(ewkt)\r\n\r\n def _db_wrangler(self, field):\r\n return pw.fn.ST_AsText(field).alias(field.name)\r\n```\r\n\r\n```\r\nquery = PadsLocations.select(PadsLocations)\r\nfor r_i, r in enumerate(query._returning):\r\n if hasattr(r, '_db_wrangler'):\r\n query._returning[r_i] = r._db_wrangler(r)\r\n```", "@davidbernat Thanks for your post - I finally got lat/lon in and out of a PeeWee PostGIS database (which I barely understand) with your code. I clearly still don't understand something, though: `ST_GeomFromText` returns `Geometry`, not `Point`, and I only get your code to work by specifying `field_type = 'Geometry'`. " ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/758/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "SarahFowler73", "id": 5266362, "node_id": "MDQ6VXNlcjUyNjYzNjI=", "avatar_url": "https://avatars.githubusercontent.com/u/5266362?v=4", "gravatar_id": "", "url": "https://api.github.com/users/SarahFowler73", "html_url": "https://github.com/SarahFowler73", "followers_url": "https://api.github.com/users/SarahFowler73/followers", "following_url": "https://api.github.com/users/SarahFowler73/following{/other_user}", "gists_url": "https://api.github.com/users/SarahFowler73/gists{/gist_id}", "starred_url": "https://api.github.com/users/SarahFowler73/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/SarahFowler73/subscriptions", "organizations_url": "https://api.github.com/users/SarahFowler73/orgs", "repos_url": "https://api.github.com/users/SarahFowler73/repos", "events_url": "https://api.github.com/users/SarahFowler73/events{/privacy}", "received_events_url": "https://api.github.com/users/SarahFowler73/received_events", "type": "User", "site_admin": false }

[]

[]

[]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/757/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/757", "html_url": "https://github.com/coleifer/peewee/pull/757", "diff_url": "https://github.com/coleifer/peewee/pull/757.diff", "patch_url": "https://github.com/coleifer/peewee/pull/757.patch", "merged_at": null }

{ "login": "therealchalz", "id": 3590449, "node_id": "MDQ6VXNlcjM1OTA0NDk=", "avatar_url": "https://avatars.githubusercontent.com/u/3590449?v=4", "gravatar_id": "", "url": "https://api.github.com/users/therealchalz", "html_url": "https://github.com/therealchalz", "followers_url": "https://api.github.com/users/therealchalz/followers", "following_url": "https://api.github.com/users/therealchalz/following{/other_user}", "gists_url": "https://api.github.com/users/therealchalz/gists{/gist_id}", "starred_url": "https://api.github.com/users/therealchalz/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/therealchalz/subscriptions", "organizations_url": "https://api.github.com/users/therealchalz/orgs", "repos_url": "https://api.github.com/users/therealchalz/repos", "events_url": "https://api.github.com/users/therealchalz/events{/privacy}", "received_events_url": "https://api.github.com/users/therealchalz/received_events", "type": "User", "site_admin": false }

[]

[]

[]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/756/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/756", "html_url": "https://github.com/coleifer/peewee/pull/756", "diff_url": "https://github.com/coleifer/peewee/pull/756.diff", "patch_url": "https://github.com/coleifer/peewee/pull/756.patch", "merged_at": null }

{ "login": "franciscod", "id": 726447, "node_id": "MDQ6VXNlcjcyNjQ0Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/726447?v=4", "gravatar_id": "", "url": "https://api.github.com/users/franciscod", "html_url": "https://github.com/franciscod", "followers_url": "https://api.github.com/users/franciscod/followers", "following_url": "https://api.github.com/users/franciscod/following{/other_user}", "gists_url": "https://api.github.com/users/franciscod/gists{/gist_id}", "starred_url": "https://api.github.com/users/franciscod/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/franciscod/subscriptions", "organizations_url": "https://api.github.com/users/franciscod/orgs", "repos_url": "https://api.github.com/users/franciscod/repos", "events_url": "https://api.github.com/users/franciscod/events{/privacy}", "received_events_url": "https://api.github.com/users/franciscod/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Is fields the right ref?\n", "No sir! It's `extra-fields`! fixing that up, sorry!\n", "push-forced the changes, but i've not yet built the docs locally so this is untested\n", "installed sphinx, `make html`, links are working! should I check anything else?\n", "Thanks a lot!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/755/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/755", "html_url": "https://github.com/coleifer/peewee/pull/755", "diff_url": "https://github.com/coleifer/peewee/pull/755.diff", "patch_url": "https://github.com/coleifer/peewee/pull/755.patch", "merged_at": "2015-11-14T13:51:34" }

{ "login": "franciscod", "id": 726447, "node_id": "MDQ6VXNlcjcyNjQ0Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/726447?v=4", "gravatar_id": "", "url": "https://api.github.com/users/franciscod", "html_url": "https://github.com/franciscod", "followers_url": "https://api.github.com/users/franciscod/followers", "following_url": "https://api.github.com/users/franciscod/following{/other_user}", "gists_url": "https://api.github.com/users/franciscod/gists{/gist_id}", "starred_url": "https://api.github.com/users/franciscod/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/franciscod/subscriptions", "organizations_url": "https://api.github.com/users/franciscod/orgs", "repos_url": "https://api.github.com/users/franciscod/repos", "events_url": "https://api.github.com/users/franciscod/events{/privacy}", "received_events_url": "https://api.github.com/users/franciscod/received_events", "type": "User", "site_admin": false }

[]

[]

[ "(edited the title so model names are general, mines are `TelegramChat` and `TwitterUser`)\n", "I'm stupid:\n\n> We still need a junction table to store the relationships between students and courses. This model can be accessed by calling the get_through_model() method. This is useful when creating tables.\n\nhttp://docs.peewee-orm.com/en/latest/peewee/playhouse.html#ManyToManyField\n\nsearch for \"creating tables\"\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/754/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "yangbh", "id": 7778764, "node_id": "MDQ6VXNlcjc3Nzg3NjQ=", "avatar_url": "https://avatars.githubusercontent.com/u/7778764?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yangbh", "html_url": "https://github.com/yangbh", "followers_url": "https://api.github.com/users/yangbh/followers", "following_url": "https://api.github.com/users/yangbh/following{/other_user}", "gists_url": "https://api.github.com/users/yangbh/gists{/gist_id}", "starred_url": "https://api.github.com/users/yangbh/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yangbh/subscriptions", "organizations_url": "https://api.github.com/users/yangbh/orgs", "repos_url": "https://api.github.com/users/yangbh/repos", "events_url": "https://api.github.com/users/yangbh/events{/privacy}", "received_events_url": "https://api.github.com/users/yangbh/received_events", "type": "User", "site_admin": false }

[]

[]

[ "here is my poor solution\n\n```\nq=basicinfo_ip_finger.insert_many(result).sql()\ndb.execute_sql(q[0]+' on duplicate key update hostname=values(hostname),...', q[1])\n```\n\nany good idea?\nthanks\n", "I looked at the docs and it seems like MySQL supports a command `REPLACE` that does what you want. I've added support for this in 2d2df5ca3b69073f4b9efef1ed6d2ad21406298b . I think this is actually preferable to using ON DUPLICATE KEY UPDATE because REPLACE does not require a separate clause describing the values.\n", "Hi there, quick question regarding this comment\r\n\r\nhttps://stackoverflow.com/a/9168948/2774823\r\n\r\nit seems sometimes ON DUPLICATE KEY UPDATE is preferable?", "Peewee does support ON DUPLICATE KEY UPDATE:\r\n\r\nhttp://docs.peewee-orm.com/en/latest/peewee/querying.html#upsert\r\n\r\n> MySQL supports upsert via the ON DUPLICATE KEY UPDATE clause. For example:\r\n\r\n```python\r\nclass User(Model):\r\n username = TextField(unique=True)\r\n last_login = DateTimeField(null=True)\r\n login_count = IntegerField()\r\n\r\n# Insert a new user.\r\nUser.create(username='huey', login_count=0)\r\n\r\n# Simulate the user logging in. The login count and timestamp will be\r\n# either created or updated correctly.\r\nnow = datetime.now()\r\nrowid = (User\r\n .insert(username='huey', last_login=now, login_count=1)\r\n .on_conflict(\r\n preserve=[User.last_login], # Use the value we would have inserted.\r\n update={User.login_count: User.login_count + 1})\r\n .execute())\r\n```", "@coleifer \r\nif I have a filed with ddl like this \r\n\r\n> `created_on` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP\r\n\r\nif I use replace, it will result in the filed \"created_on\" is \"updated\" to current time, because the origin record is deleted and a new record created,this is not my expected.Otherwise, Insert into ..... ON DUPLICATE KEY UPDATE work well.\r\n", "@xzycn - you should consult the documentation for your database, my dude. It clearly says that REPLACE works by deleting and re-inserting in the event of a conflict.\r\n\r\nhttps://dev.mysql.com/doc/refman/8.0/en/replace.html", "@coleifer \r\nI know that, just didn't claim it.I just want to say:\"it seems sometimes ON DUPLICATE KEY UPDATE is preferable\", thank you for your reply :)" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/753/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "bcthomas", "id": 86821, "node_id": "MDQ6VXNlcjg2ODIx", "avatar_url": "https://avatars.githubusercontent.com/u/86821?v=4", "gravatar_id": "", "url": "https://api.github.com/users/bcthomas", "html_url": "https://github.com/bcthomas", "followers_url": "https://api.github.com/users/bcthomas/followers", "following_url": "https://api.github.com/users/bcthomas/following{/other_user}", "gists_url": "https://api.github.com/users/bcthomas/gists{/gist_id}", "starred_url": "https://api.github.com/users/bcthomas/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/bcthomas/subscriptions", "organizations_url": "https://api.github.com/users/bcthomas/orgs", "repos_url": "https://api.github.com/users/bcthomas/repos", "events_url": "https://api.github.com/users/bcthomas/events{/privacy}", "received_events_url": "https://api.github.com/users/bcthomas/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You can use the `Proxy` object [see circular foreign keys](http://docs.peewee-orm.com/en/latest/peewee/models.html#circular-foreign-key-dependencies).\n\n``` python\nclass User(BaseModel):\n email = CharField(unique=True)\n\nProjectUserProxy = Proxy()\n\nclass Project(BaseModel):\n name = CharField(null=True)\n users = ManyToManyField(\n User, \n related_name='projects', \n through_model=ProjectUserProxy)\n\n class Meta:\n db_table = 'projects'\n\nclass ProjectUser(BaseModel):\n project = ForeignKeyField(Project, db_column='project_id', index=True, null=True)\n user = ForeignKeyField(User, db_column='user_id', index=True, null=True)\n\n class Meta:\n db_table = 'project_users'\n\nProjectUserProxy.initialize(ProjectUser)\n```\n\nEverything worked correctly with the above code when I tried creating some users, projects, and the relationships between them. The custom through table appeared to be working as well.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/752/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "kramer65", "id": 596581, "node_id": "MDQ6VXNlcjU5NjU4MQ==", "avatar_url": "https://avatars.githubusercontent.com/u/596581?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kramer65", "html_url": "https://github.com/kramer65", "followers_url": "https://api.github.com/users/kramer65/followers", "following_url": "https://api.github.com/users/kramer65/following{/other_user}", "gists_url": "https://api.github.com/users/kramer65/gists{/gist_id}", "starred_url": "https://api.github.com/users/kramer65/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/kramer65/subscriptions", "organizations_url": "https://api.github.com/users/kramer65/orgs", "repos_url": "https://api.github.com/users/kramer65/repos", "events_url": "https://api.github.com/users/kramer65/events{/privacy}", "received_events_url": "https://api.github.com/users/kramer65/received_events", "type": "User", "site_admin": false }

[]

[]

[ "The problem, not obvious at all, is that the `ticket` column is not found. I bet that if you change it to `ticket_id` you will be set:\n\n``` python\nfrom playhouse.migrate import MySQLMigrator, migrate\nmigrator = MySQLMigrator(db.database)\nmigrate(migrator.drop_not_null('emailsending', 'ticket_id'))\n```\n", "Looking at the tests, there may be some additional issues with the mysql migrations regarding dropping constraints on foreign keys. Looking into it.\n", "@coleifer - Seeing that you closed this issue I presume that after you looked into it you concluded that there were no additional issues with the mysql migrations regarding dropping constraints on foreign keys?\n\nThe reason I'm asking is that I'm using this in beta/production now and I wanted to wait with migrating the db until you had a conclusion.\n", "Just to be safe I added another testcase adding and dropping not null constraints to a foreign key to verify they worked.\n", "Thanks for adding the tests. I tested it on 3 test setups before deploying the migration to the production server, and everything went perfectly fine.\n\nThanks for creating peewee!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/751/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "arpankapoor", "id": 3677810, "node_id": "MDQ6VXNlcjM2Nzc4MTA=", "avatar_url": "https://avatars.githubusercontent.com/u/3677810?v=4", "gravatar_id": "", "url": "https://api.github.com/users/arpankapoor", "html_url": "https://github.com/arpankapoor", "followers_url": "https://api.github.com/users/arpankapoor/followers", "following_url": "https://api.github.com/users/arpankapoor/following{/other_user}", "gists_url": "https://api.github.com/users/arpankapoor/gists{/gist_id}", "starred_url": "https://api.github.com/users/arpankapoor/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/arpankapoor/subscriptions", "organizations_url": "https://api.github.com/users/arpankapoor/orgs", "repos_url": "https://api.github.com/users/arpankapoor/repos", "events_url": "https://api.github.com/users/arpankapoor/events{/privacy}", "received_events_url": "https://api.github.com/users/arpankapoor/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Does the response middleware method not need to return anything? (Or the request middleware for that matter)\n", "Nothing needs to be returned.\n\nThe middleware methods may modify the request and response objects passed in as parameters. These objects are routed to the appropriate responder method by Falcon.\n\nSee the docstring of the middleware methods [here](https://falcon.readthedocs.org/en/latest/api/middleware.html#middleware-components). \n", "Thanks!\n", "Not sure if it should be documented, but calling db.close in the process_response kills the cursor, and then a response via stream will fail\r\n\r\nA poor man workaround in this case could be to check the existence of the stream:\r\n\r\n```\r\n def process_response(self, req, resp, resource):\r\n if resp.stream==None and not db.is_closed():\r\n db.close()\r\n```\r\n\r\nAnd then use some stream incorporating a close() method, which falcon will call.\r\n\r\n```\r\nclass closingMap(map):\r\n def close(self):\r\n db.close()\r\n```\r\n for python > 3.6 some technique more \"awaitable\" could be better\r\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/750/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/750", "html_url": "https://github.com/coleifer/peewee/pull/750", "diff_url": "https://github.com/coleifer/peewee/pull/750.diff", "patch_url": "https://github.com/coleifer/peewee/pull/750.patch", "merged_at": "2015-11-06T20:27:49" }

{ "login": "yohanboniface", "id": 146023, "node_id": "MDQ6VXNlcjE0NjAyMw==", "avatar_url": "https://avatars.githubusercontent.com/u/146023?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yohanboniface", "html_url": "https://github.com/yohanboniface", "followers_url": "https://api.github.com/users/yohanboniface/followers", "following_url": "https://api.github.com/users/yohanboniface/following{/other_user}", "gists_url": "https://api.github.com/users/yohanboniface/gists{/gist_id}", "starred_url": "https://api.github.com/users/yohanboniface/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/yohanboniface/subscriptions", "organizations_url": "https://api.github.com/users/yohanboniface/orgs", "repos_url": "https://api.github.com/users/yohanboniface/repos", "events_url": "https://api.github.com/users/yohanboniface/events{/privacy}", "received_events_url": "https://api.github.com/users/yohanboniface/received_events", "type": "User", "site_admin": false }

[]

[]

[ "You need to use `.filter(one__is=None)`\n", "To be more specific, the filter API has been deprecated since 2.0.0. Why are you using it?\n", "> You need to use .filter(one__is=None)\n\nAh, ok, thanks.\n\n> To be more specific, the filter API has been deprecated since 2.0.0. Why are you using it?\n- Easier to build dynamically (for example from query parameters (?xxx=yyy&aaa=bbb), or any situation where you need to map from key/value).\n- More concise (we need to fit in 80 lines, no? ;) ).\n- more consistant: `Class.field != value` vs `Class.field.contains(value)`\n- feels pythonic at first view, but then we hit the \"or\", \"not\", \"in\"… limit, so a bit frustrating system at use _imho_\n- my brain refuses to remember `<<` and `>>` own operation ;)\n\nCan you explain (or point me to an explanation about) why it is deprecated? :)\n", "Well, you don't have to use `<<` and `>>`. You can use:\n\n``` python\nClass.field.in_()\nClass.field.is_null(True)\nClass.field.is_null(False)\n```\n\nAs to being concise, the way to format your code is up to you of course, but I never have issues with 80 character lines, e.g.\n\n``` python\n# Get blogs tagged with \"python\".\npython_blogs = (BlogEntry\n .select()\n .join(BlogEntryTags)\n .join(Tag)\n .where(\n (BlogEntry.is_published == True) &\n (Tag.tag == 'python'))\n .order_by(BlogEntry.timestamp.desc()))\n```\n\n---\n\nThe APIs are deprecated because they are not good APIs. Keyword arguments do not preserve ordering, which can be important when constructing where clauses that will hit an index. Keyword arguments necessitate hacks like `DQ(foo__bar='whatev') | DQ(whatever='nug')`. The set of operators is not nearly as rich. You cannot easily do things like express SQL aggregations, function calls, etc.\n", "> Keyword arguments do not preserve ordering, which can be important when constructing where clauses that will hit an index.\n\nSounds like a strong argument.\nThanks your answer. :)\n", "What about adding a [DeprecationWarning](https://docs.python.org/2/library/exceptions.html#exceptions.DeprecationWarning) when the function is called then?\n", "> To be more specific, the filter API has been deprecated since 2.0.0.\n\nDoes that means `get_or_create` is deprecated too, given that it relies on `filter`? If yes, is there a recommended way (bonus: that allows using HStore fields)?\n", "`get_or_create()` is not recommended, but is left in place because a lot of people have come to rely on it. I never use it myself, preferring to code things up explicitly:\n\n``` python\nwith db.atomic():\n try:\n user = User.create(username='whatever') # unique constraint on username\n except IntegrityError:\n user = User.get(User.username == 'whatever')\n```\n", "Thanks for the suggestion, that's indeed a nice pattern, but that's a `create_or_get`, which works only when you have a DB constraint to rely on. Anyway, I'll will dig this way. Thanks!\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/749/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "notdaniel", "id": 3722519, "node_id": "MDQ6VXNlcjM3MjI1MTk=", "avatar_url": "https://avatars.githubusercontent.com/u/3722519?v=4", "gravatar_id": "", "url": "https://api.github.com/users/notdaniel", "html_url": "https://github.com/notdaniel", "followers_url": "https://api.github.com/users/notdaniel/followers", "following_url": "https://api.github.com/users/notdaniel/following{/other_user}", "gists_url": "https://api.github.com/users/notdaniel/gists{/gist_id}", "starred_url": "https://api.github.com/users/notdaniel/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/notdaniel/subscriptions", "organizations_url": "https://api.github.com/users/notdaniel/orgs", "repos_url": "https://api.github.com/users/notdaniel/repos", "events_url": "https://api.github.com/users/notdaniel/events{/privacy}", "received_events_url": "https://api.github.com/users/notdaniel/received_events", "type": "User", "site_admin": false }

[]

[]

[ "What database are you using?\n", "Ha! I was using SQLite (which I usually use when starting something out) but after your response, I just tried it with Postgres and it worked fine. Is this a SQLite limitation, then? Either way, this is helpful since I was going to move this to Postgres eventually anyway.\n", "SQLite does not support adding constraints. So basically you can follow the documentation example and simply skip the call to add the constraint. I've added a test and updated the docs to make this clear.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/748/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "josefdlange", "id": 1062835, "node_id": "MDQ6VXNlcjEwNjI4MzU=", "avatar_url": "https://avatars.githubusercontent.com/u/1062835?v=4", "gravatar_id": "", "url": "https://api.github.com/users/josefdlange", "html_url": "https://github.com/josefdlange", "followers_url": "https://api.github.com/users/josefdlange/followers", "following_url": "https://api.github.com/users/josefdlange/following{/other_user}", "gists_url": "https://api.github.com/users/josefdlange/gists{/gist_id}", "starred_url": "https://api.github.com/users/josefdlange/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/josefdlange/subscriptions", "organizations_url": "https://api.github.com/users/josefdlange/orgs", "repos_url": "https://api.github.com/users/josefdlange/repos", "events_url": "https://api.github.com/users/josefdlange/events{/privacy}", "received_events_url": "https://api.github.com/users/josefdlange/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Immediately after posting it I of course thought of the way a method or property lives on the class level, and how it might be difficult to reconcile to an instance's bound method. Hmm...\n", "I'm seeing the difficulty here. If you pass in the unbound method, there's no (apparent) way to match a bound instance method with said unbound class method. If there's no way you can figure out either, I'll probably close this one up.\n", "@coleifer Which means itsn't possible to get `extra_attrs` on recursive items?", "No, for such cases as that you're probably best off just writing your own serializer." ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/747/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "josefdlange", "id": 1062835, "node_id": "MDQ6VXNlcjEwNjI4MzU=", "avatar_url": "https://avatars.githubusercontent.com/u/1062835?v=4", "gravatar_id": "", "url": "https://api.github.com/users/josefdlange", "html_url": "https://github.com/josefdlange", "followers_url": "https://api.github.com/users/josefdlange/followers", "following_url": "https://api.github.com/users/josefdlange/following{/other_user}", "gists_url": "https://api.github.com/users/josefdlange/gists{/gist_id}", "starred_url": "https://api.github.com/users/josefdlange/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/josefdlange/subscriptions", "organizations_url": "https://api.github.com/users/josefdlange/orgs", "repos_url": "https://api.github.com/users/josefdlange/repos", "events_url": "https://api.github.com/users/josefdlange/events{/privacy}", "received_events_url": "https://api.github.com/users/josefdlange/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I am going to pass on this particular API or implementation. The `model_to_dict` is a simple but flexible piece of code, but it mostly stands on its own. Introducing APIs at the model level seems like too tight a coupling.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/746/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "scribu", "id": 225715, "node_id": "MDQ6VXNlcjIyNTcxNQ==", "avatar_url": "https://avatars.githubusercontent.com/u/225715?v=4", "gravatar_id": "", "url": "https://api.github.com/users/scribu", "html_url": "https://github.com/scribu", "followers_url": "https://api.github.com/users/scribu/followers", "following_url": "https://api.github.com/users/scribu/following{/other_user}", "gists_url": "https://api.github.com/users/scribu/gists{/gist_id}", "starred_url": "https://api.github.com/users/scribu/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/scribu/subscriptions", "organizations_url": "https://api.github.com/users/scribu/orgs", "repos_url": "https://api.github.com/users/scribu/repos", "events_url": "https://api.github.com/users/scribu/events{/privacy}", "received_events_url": "https://api.github.com/users/scribu/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Postgres is actually right in this case, and this isn't a bug in peewee. It looks like the value returned by the attribute lookup is not numeric. You can cast it, but you have to cast both return values it seems.\n\nOne way that seemed to work was:\n\n``` sql\nSELECT \nCASE \"t1\".\"price_info\"->>'currency' \nWHEN 'EUR' THEN (\"t1\".\"price_info\"->>'value')::float \nWHEN 'RON' THEN ((\"t1\".\"price_info\"->>'value')::float * 1.5) \nEND AS price \nFROM \"xxtest\" AS t1;\n```\n\nTo accomplish this with peewee, you can do something like this. It's not super elegant, but it should work.\n\n``` python\nfrom peewee import Clause, SQL, fn\n\ndef to_float(node):\n return fn.CAST(Clause(node, SQL('AS float')))\n\nprice = case(\n Posting.price_info['currency'],\n (\n ('EUR', to_float(Posting.price_info['value'])),\n ('RON', to_float(Posting.price_info['value']) * 1.5),\n )\n).alias('price')\n```\n", "With 73cfdc1824f4161db062fc74fc77a5a13660888e it should be possible to now write:\n\n```\nprice = case(\n Posting.price_info['currency'],\n (\n ('EUR', Posting.price_info['value'].cast('float')),\n ('RON', Posting.price_info['value'].cast('float') / ONE_EURO_IN_RON)\n )\n).alias('price')\n```\n", "Works! Thanks for the syntax sugar.\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/745/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "login": "franciscod", "id": 726447, "node_id": "MDQ6VXNlcjcyNjQ0Nw==", "avatar_url": "https://avatars.githubusercontent.com/u/726447?v=4", "gravatar_id": "", "url": "https://api.github.com/users/franciscod", "html_url": "https://github.com/franciscod", "followers_url": "https://api.github.com/users/franciscod/followers", "following_url": "https://api.github.com/users/franciscod/following{/other_user}", "gists_url": "https://api.github.com/users/franciscod/gists{/gist_id}", "starred_url": "https://api.github.com/users/franciscod/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/franciscod/subscriptions", "organizations_url": "https://api.github.com/users/franciscod/orgs", "repos_url": "https://api.github.com/users/franciscod/repos", "events_url": "https://api.github.com/users/franciscod/events{/privacy}", "received_events_url": "https://api.github.com/users/franciscod/received_events", "type": "User", "site_admin": false }

[]

[]

[ "Thank you!\n", "np! :)\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/744/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/744", "html_url": "https://github.com/coleifer/peewee/pull/744", "diff_url": "https://github.com/coleifer/peewee/pull/744.diff", "patch_url": "https://github.com/coleifer/peewee/pull/744.patch", "merged_at": "2015-11-02T14:51:03" }

{ "login": "Maillol", "id": 2394534, "node_id": "MDQ6VXNlcjIzOTQ1MzQ=", "avatar_url": "https://avatars.githubusercontent.com/u/2394534?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Maillol", "html_url": "https://github.com/Maillol", "followers_url": "https://api.github.com/users/Maillol/followers", "following_url": "https://api.github.com/users/Maillol/following{/other_user}", "gists_url": "https://api.github.com/users/Maillol/gists{/gist_id}", "starred_url": "https://api.github.com/users/Maillol/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/Maillol/subscriptions", "organizations_url": "https://api.github.com/users/Maillol/orgs", "repos_url": "https://api.github.com/users/Maillol/repos", "events_url": "https://api.github.com/users/Maillol/events{/privacy}", "received_events_url": "https://api.github.com/users/Maillol/received_events", "type": "User", "site_admin": false }

[]

[]

[ "I've waffled about this some. I dislike implicitly creating the columns. Additionally, when querying, things seem like they might get a little weird, e.g.:\n\n``` python\nclass Person(Model):\n first_name = CharField()\n last_name = CharField()\n class Meta:\n primary_key = CompositeKey('first_name', 'last_name')\n\n# API from your PR:\nclass Pet(Model):\n owner = ForeignKeyField(Person)\n\n# API I had considered:\nclass Pet(Model):\n owner_first_name = CharField()\n owner_last_name = CharField()\n owner = CompositeForeignKey(Person)\n```\n\nEach has its downsides. In the former, the related person's first and last name, which are present on the Pet, are abstracted away and not easily accessible. In the latter, I'm left wondering how useful that `owner` attribute is really going to be.\n", "I agree with you, creating columns must be explicit.\n\ncurrently the problem appear when we want create a weak entity which contains an other weak entity. \ni.e:\nBook -> Chapter -> Paragraph \nor\nTrain -> Wagon -> Seat\n\nThe PK Wagon is (PK train + num Wagon) and PK Seat is (PK train + num Wagon + num Seat).\nThe Wagon num field isn't unique, I can't create a foreign key on it, I must use composite FK. \n\nWhat do you think about API like this ?\n\n``` python\nclass Pet(Model):\n owner_first_name = ForeignKey(Person, first_name)\n owner_last_name = ForeignKey(Person, last_name)\n\n class Meta:\n foreign_keys = [\n CompositeKey(owner_first_name, owner_last_name)\n ]\n```\n", "> What do you think about API like this ?\n\nThat would be confusing I think. By making `owner_first_name` and `owner_last_name` ForeignKeyFields, it is unclear what value you are supposed to assign to them. Typically with a foreign key field you would assign the related model instance (although it is possible to assign the underyling scalar key value). This \"overloads\" the concept of foreign keys in a way I find ambiguous.\n\nIf there were going to be an API, it would be this most likely:\n\n``` python\nclass Pet(Model):\n owner_first_name = CharField()\n owner_last_name = CharField()\n owner = CompositeForeignKey(Person)\n```\n\nBut as I said I have doubts about the utility of this.\n\nI think that for now I am going to close this as _wontfix_, since there doesn't seem to me to be a clear path forward. You can specify table-level constraints as model meta options, so if you want schema-level referential integrity, you can do something like:\n\n``` python\nclass Pet(Model):\n owner_first_name = CharField()\n owner_last_name = CharField()\n\n class Meta:\n constraints = [SQL('FOREIGN KEY(owner_first, owner_last) REFERENCES person(first, last)')]\n```\n", "Documentation: http://docs.peewee-orm.com/en/latest/peewee/models.html#indexes-and-constraints\n" ]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/743/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/743", "html_url": "https://github.com/coleifer/peewee/pull/743", "diff_url": "https://github.com/coleifer/peewee/pull/743.diff", "patch_url": "https://github.com/coleifer/peewee/pull/743.patch", "merged_at": null }

{ "login": "bakemecookies", "id": 2167033, "node_id": "MDQ6VXNlcjIxNjcwMzM=", "avatar_url": "https://avatars.githubusercontent.com/u/2167033?v=4", "gravatar_id": "", "url": "https://api.github.com/users/bakemecookies", "html_url": "https://github.com/bakemecookies", "followers_url": "https://api.github.com/users/bakemecookies/followers", "following_url": "https://api.github.com/users/bakemecookies/following{/other_user}", "gists_url": "https://api.github.com/users/bakemecookies/gists{/gist_id}", "starred_url": "https://api.github.com/users/bakemecookies/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/bakemecookies/subscriptions", "organizations_url": "https://api.github.com/users/bakemecookies/orgs", "repos_url": "https://api.github.com/users/bakemecookies/repos", "events_url": "https://api.github.com/users/bakemecookies/events{/privacy}", "received_events_url": "https://api.github.com/users/bakemecookies/received_events", "type": "User", "site_admin": false }

[]

[]

[]

{ "url": "https://api.github.com/repos/coleifer/peewee/issues/742/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0 }

{ "url": "https://api.github.com/repos/coleifer/peewee/pulls/742", "html_url": "https://github.com/coleifer/peewee/pull/742", "diff_url": "https://github.com/coleifer/peewee/pull/742.diff", "patch_url": "https://github.com/coleifer/peewee/pull/742.patch", "merged_at": "2015-11-01T02:16:10" }