Calling list() on an empty sliced union still causes a database query

[Lucidiot]

When upgrading from Django 4.1 to 4.2, some of our unit tests failed because of unexpected SQL queries, and those queries are still present in Django 5.0.

We have a Django REST Framework API view, which relies on Django's Paginator to do its pagination. It uses a QuerySet which involves a UNION of two other queries that both have some filters. Those filters can cause the whole union to be empty. The Paginator knows that the result count is 0, so it returns a page of results by slicing with [0:0].

In Django 4.1, calling list() on that sliced union would cause no query and an empty list is returned, but starting with 4.2, a query runs:

qs = Thing.objects.filter(id__in=Thing.objects.none())
list(qs)                 # 0 queries
list(qs[0:0])            # 0 queries
list(qs.union(qs))       # 0 queries
list(qs.union(qs)[0:0])  # 1 query

(
    SELECT "app_thing"."id" AS "col1"
    FROM "app_thing"
    WHERE 0 = 1
) UNION (
    SELECT "app_thing"."id" AS "col1"
    FROM "app_thing"
    WHERE 0 = 1
)

I think this extra query was added in the fix for #34125.

[Simon Charette]

I confirm that c2cc80756b8949cdd87b88bbfdfee698ced441e0 focused on the correctness aspect of the problem (avoiding the crash) without focusing on some edge cases that might still allow the query eliding to take place.

[Simon Charette]

Wrong assessment, test coverage exists as the crash detailed in #34125 that this ticket change the approach for (to prevent unnecessary queries) only manifests itself on Postgres.

[Sarah Boyce <42296566+sarahboyce@…>]

In 9cb8baa0:

Fixed #35559 -- Avoided unnecessary query on sliced union of empty queries.

While refs #34125 focused on the SQL correctness of slicing of union of
potentially empty queries it missed an optimization opportunity to avoid
performing a query at all when all queries are empty.

Thanks Lucidiot for the report.