List and Count Method on Respository

[edimedia]

Hi! I want to know why the authors chose to make a count request to the DB instead of using len(instances).

https://github.com/litestar-org/advanced-alchemy/blob/24f7237ee024aa531ac54507cc66f5786245231f/advanced_alchemy/repository/_async.py#L1081

        count = count_result.scalar_one()
        result = await self._execute(statement, loader_options_have_wildcards=loader_options_have_wildcard)
        instances: list[ModelT] = []
        for (instance,) in result:
            self._expunge(instance, auto_expunge=auto_expunge)
            instances.append(instance)
        return instances, count

instead of

        result = await self._execute(statement, loader_options_have_wildcards=loader_options_have_wildcard)
        instances: list[ModelT] = []
        for (instance,) in result:
            self._expunge(instance, auto_expunge=auto_expunge)
            instances.append(instance)
        return instances, len(instances)          # <=========

I apologize if this question is not appropriate for this context.

[cofin]

@edimedia, the list_and_count method returns a tuple of the rows you selected and the "total rows" available. This is typically used in places where you want to do pagination.

Consider a case where you have passed in a PaginationFilter. In this case, you want your data to be the limited result set, however, you need the total row count in most cases to determine your "current page". In short we return (<query results>, <count of results without any limit filters applied>)

Depending on the database engine you are using, Advanced Alchemy will either add in the row count with a analytical window statement. This lets us get the limited rows and the total count in a single query. Otherwise, it issues 2 queries. 1 for the total count, and 1 for the data requested.

Hope this helps.

[edimedia]

Thanks, that's clear !