Right?

if you remove '+00:00' if will be parsed by Pony as timezone-naive datetime value

Pony expects all datetime values are timezone-naive in UTC

Hmm

I'll try that

works

thank you

Sure

Hey, I got another question. Does async work with sqlite?

Hi! Async support in Pony is not working properly. Are your application async?

Yes

Yes

It seems that Pony has something for working with the database right from the front.

It's basically async functions that get executed when an event occurs

I wonder if it blocks the thread if I use it just like that

It seems that Pony has something for working with the database right from the front.

I didn't understand

I saw something like that, but I remember it vaguely.

It seems that Pony has something for working with the database right from the front.

There was PonyJS library, but it wasn't officially released

I wonder if it blocks the thread if I use it just like that

You can use db_session inside async functions. But such functions should not do async calls and context switches inside db_session. If you want to do some complex processing which can take a lot of time it is better to do it in a separate thread

I'm not very experienced, just to make sure I understood. So if I use await in the same function, it is a problem?

You can use await in the same function, but not inside db_session

I see

Thank you

I'm not very experienced, just to make sure I understood. So if I use await in the same function, it is a problem?

maybe use db_session as a context manager?

You mean with with?

Yes

Instead of the annotation

Yeah, I'll do that

Thanks guys

It's should be possible to offload the db call into another thread through

and make it async that way

loop = asyncio.get_running_loop() with concurrent.futures.ThreadPoolExecutor() as pool: result = await loop.run_in_executor(pool, functools.partial(yoursyncfunction, args))

Thank you I'll try that

Good that you actually included some code snippet xD

is there any simple way to explode a list as parameter(to using in a "in") when using Database.select?

If you do mylist = [1, 2, 3] query = select(x for x in X if x.id in mylist) it should work

sorry, i refering using raw queries ... db.select("select * from table where field in ($list)")

sorry, i refering using raw queries ... db.select("select * from table where field in ($list)")

You can construct it manually mylist = [1, 2, 3] sql = """ select * from t1 where field in (%s) ""' % ', '.join("$mylist[%d]" % i for i in range(mylist)) db.select(sql)

thx

I'm working on adding support for SQL Server, any suggestion which of the tests I should test it with? I've looked through a few but most of them seem related to pony itself rather than testing the dbproviders

Any recomendation to convert a <class 'pony.orm.core.row'> into a Dict ?

pony.orm.core.DBSchemaError: Foreign key fk_facebookcommentmoderationaction_facebookcommentmoderationrul cannot be created, name is already in use I think I hit this before, could it be a length issue? that’s 63 chars long

How can I set a custom index name? I can’t see how in the docs

How can I set a custom index name? I can’t see how in the docs

You can specify attribute options index='some_index_name' and fk_name='some_fk_name'

Thank you 🙂

https://docs.ponyorm.org/api_reference.html?highlight=fk_name#cmdoption-arg-fk_name

fk_name worked

Maybe the error message could be better?

This error is generated by database. Pony need to deal with very long names that exceeds max name length limit by cutting it and adding some hash value to the end, it will be added with migrations

sounds good!

Any recomendation to convert a <class 'pony.orm.core.row'> into a Dict ?

At this moment there is no universal way to do this. Note that you can access row columns by name, if the column name is valid Python identifier: rows = db.select("* from student") for row in rows: print(row[0], row[1]) print(row.name, row.gpa) To convert row to dict you need to know query result's column names: names = ['name', 'gpa'] d = {name: row[i] for i, name in enumerate(names)} Maybe we should add some generic method for converting a row to a dict

I'm working on adding support for SQL Server, any suggestion which of the tests I should test it with? I've looked through a few but most of them seem related to pony itself rather than testing the dbproviders

The tests for database syntax are not separated from tests for pony itself, I think you need to run entire test suite: cd pony/pony/orm/tests python -m unittest discover It is not easy to separate tests, because for some databases logic of Pony itself should be changed to satisfy database details. For example, in Oracle NULL and empty string are the same thing, whereas in other databases they are different

Are there any plans to add async support to pony? (With the @db_session decorator)

Not in the near time, as async is not easily compatible with IdentityMap pattern used by Pony

I see. Thanks!

Hi: class AAA(Entity): bbbs = Set("B") class BBB(Entity): aaa = REquired(A)is it possible to retrieve "bbbs" (I mean the name of the attribute") but from the BBB class ?

Yes: BBB.aaa.reverse.name

thanks, is it possible to put some attibute (Required, Optional) inside a Mixin ? I get an assertion error : def load(attr, obj): cache = obj._session_cache_ if cache is None or not cache.is_alive: throw_db_session_is_over('load attribute', obj, attr) if not attr.columns: reverse = attr.reverse > assert reverse is not None and reverse.columns E AssertionError

Mixins are not supported yet

ok

Hi

Just got this error after migrating my mariadb to pg

TypeError: Attribute table.col: PostgreSQL provider does not support unsigned bigint type

=\

col = Required(int, size=64, unsigned=True, default=0)

I need to speficify the size or everything won't run since pony will limit between int min and max

PostgreSQL is indeed does not support unsigned integers. When you use unsigned=True in attribute description Pony uses bigger size to be able to keep all values (say, integer instead of unsigned smallint). But there is no bigger size for bigint https://stackoverflow.com/questions/20810134/why-unsigned-integer-is-not-available-in-postgresql

Oh its pg thing, sorry for my lack of knowledge and its about unsigned, not size as I though at first

Thanks @metaprogrammer

Sure

Hey! An online diagram editing tool has the checkbox "Hidden" in entity settings, but I found nothing about it in docs

What does it do?

It was for PonyJS library which was not officially released, we need to remove it

thanks for the answer and your cool library ❤️

hi, return select( matiere for groupe in self.groupes.order_by(lambda g: g.position) for matiere in groupe.matieres.order_by(lambda m: m.position) ) position` is a property for a _position attribute. I get the following error : AttributeError: groupe.matieres.select().sort_by

return select( matiere for groupe in self.groupes.order_by(GroupeMatiere.position) for matiere in groupe.matieres.order_by(Matiere.position) ) return the error TypeError: order_by() method receive an argument of invalid type: <property object at 0x7fbba8a9c400>

and finally : return select( matiere for groupe in self.groupes.order_by(GroupeMatiere._position) for matiere in groupe.matieres.order_by(Matiere._position) ) TypeError: Expression Matiere._position has unsupported type 'Required'

Any idea on it ?

Any idea on it ?

Can you please send the model definition as well, please?

You need to apply ordering to the outer query, something like that: select( matiere for groupe in self.groupes for matiere in groupe.matieres ).order_by( lambda matiere: (matiere.groupe.position, matiere.position) )

yeah that was simply that, tahnks