Можно ли использовать PonyORM без предварительного задания моделей? Типа select(o for o in Table('MyTable'))

ORM works with defined entities in general

Define entities and try this: db.generate_mapping(create_tables=False)

Define entities and try this: db.generate_mapping(create_tables=False)

I'm not sure I want to add it to the current version of Pony. The way PonyORM deals with database connections and implements its in-memory cache of loaded objects is not too suitable for long-live async coroutines. But I'd like to implement a new version of PonyORM (with a slightly different API) with the following properties: 1. Full support of typing, so all Pony expressions are properly typed and recognized by IDEs like PyCharm (I've already tested that it is possible) 2. Async support 3. Migrations from the beginning 4. Dataclass-like syntax for model definitions 5. Integration with Pydantic 6. Easy conversion query results to JSON, including hierarchical data Realistically, I can't start work on it before the new year.

Hey I'd like to look into async as well. Maybe we can even figure out a way how we can both work on that together. Been playing around with the pydantic stuff and created my own light ORM for postgres, https://github.com/luckydonald/fastorm

Hey I'd like to look into async as well. Maybe we can even figure out a way how we can both work on that together. Been playing around with the pydantic stuff and created my own light ORM for postgres, https://github.com/luckydonald/fastorm

Hi! Yes. I should have some time from January 17th to February 11th for this, so I hope to add Python 3.12 support during this period.

any chance you already take a look Alex?

Super interested in making a donation to get Enum support

Hello, does anyone know if there is a way to change the database connection to the running application? The user can change the provider, user, password, dsn

You can disconnect and create a new connection. There should be some documentation about removing a connection in unit tests somewhere, this should work here too

Hello everyone, i have a question since there is no database migration, available right now in pony orm, how do i go about making changes after my database is already in product, because i think manually altering the database can be dangerous

Tldr: move the entity definitions inside a function

got to this, It would be like this?

I define a function for each entity, in different files

def define_entities(db): class Foo(db.Entity): ... class Bar(db.Entity): ... def open_database(filename): db = Database() define_entities(db) db.bind('sqlite', filename) db.generate_mapping(create_tables=True) return db

This is the final way I arrived to achieve the expected result

Suppose you have an entity class MyEntity(db.Entity): id = PrimaryKey(int) title = Required(str) description = Required(str) ... After you created the database and tables, you can execute an additional commands like the following: sql_create_fts_table = """ CREATE VIRTUAL TABLE IF NOT EXISTS FtsIndex USING FTS5 (title, description, content='MyEntity', tokenize='porter unicode61 remove_diacritics 1');""" sql_add_fts_trigger_insert = """ CREATE TRIGGER IF NOT EXISTS fts_ai AFTER INSERT ON MyEntity BEGIN INSERT INTO FtsIndex(rowid, title, description) VALUES (new.rowid, new.title, new.description); END;""" sql_add_fts_trigger_delete = """ CREATE TRIGGER IF NOT EXISTS fts_ad AFTER DELETE ON MyEntity BEGIN DELETE FROM FtsIndex WHERE rowid = old.rowid; END;""" sql_add_fts_trigger_update = """ CREATE TRIGGER IF NOT EXISTS fts_au AFTER UPDATE ON MyEntity BEGIN DELETE FROM FtsIndex WHERE rowid = old.rowid; INSERT INTO FtsIndex(rowid, title) VALUES (new.rowid, new.title, new.description); END;""" with db_session(ddl=True): cursor = self.db.get_connection().cursor() cursor.execute(sql_create_fts_table) cursor.execute(sql_add_fts_trigger_insert) cursor.execute(sql_add_fts_trigger_delete) cursor.execute(sql_add_fts_trigger_update) and then in your query you can do something like fts_query = "foo bar*" select( obj for obj in MyEntity if raw_sql(""" id in SELECT(rowid from FtsIndex WHERE FtsIndex MATCH $fts_query) """) ) Can't say about spatialite, I haven't had experience with it

good morning, worried about the performance of the IN (SELECT ...) part. wondering if it could be replaced by a join somehow

Hello, people! Can somebody help me to understand something? I have a Postgres behind Yandex Odyssey connection pooler. When I try to name connection through db.bind(application_name="MYAPP"), it does not work. When I set name of connection through execution db.execute('set application_name = "MYAPP";') on session start, it works fine inside the session, but when session ends and I open new session, application_name resets to nothing. I can, probably, name it in every session with custom decorator or something, but why it works this way? Isn't that true, that in one thread should be one connectiona in pool, so when I run sessions in one thread, it should be same connection? Just curious. And if someone can think of the reason why Odyssey works this way, it woud be great) Here's code example: from pony.orm import Database, db_session db = Database() db.bind( provider='postgres', user='...', password='...', host='...', database='...', target_session_attrs='read-write', application_name='THIS NAME DO NOT APPEAR', ) db.generate_mapping() with db_session(): print('session') db.execute('set application_name = "THIS NAME DO APPEAR";') cur = db.execute("SELECT application_name FROM pg_stat_activity WHERE application_name like 'THIS%'") print(cur.fetchall()) print('we see "THIS NAME DO APPEAR" in connection name list') with db_session(): print('session') cur = db.execute("SELECT application_name FROM pg_stat_activity WHERE application_name like 'THIS%'") print(cur.fetchall()) print('we do not see name in connection name list')

Will there be a migration system in pony orm?