#37140: NULL handling of exclude() with __in varies between nullable expressions
versus literal None
-------------------------------------+-------------------------------------
     Reporter:  Sebastian Vera       |                    Owner:  (none)
         Type:  Bug                  |                   Status:  new
    Component:  Database layer       |                  Version:  dev
  (models, ORM)                      |
     Severity:  Normal               |               Resolution:
     Keywords:  exclude in subquery  |             Triage Stage:
  NULL NOT IN                        |  Unreviewed
    Has patch:  0                    |      Needs documentation:  0
  Needs tests:  0                    |  Patch needs improvement:  0
Easy pickings:  0                    |                    UI/UX:  0
-------------------------------------+-------------------------------------
Comment (by Simon Charette):

 Replacing `__in` to perform `EXISTS` query is likely never going to happen
 for the reasons discussed in the thread so we'd likely need to introduce
 an `__exists` lookup instead to avoid breaking backward compatibility so
 knowing that I wouldn't be opposed to documenting this pitfall even if I
 suspect folks would only stumble upon the edge case after the fact.

 As for ''what'' should be documented as a workaround having the subquery
 explicitly exclude `NULL` (e.g. `filter(inner_field__isnull=False)`) is
 likely going to be closer to what the ORM would normally generate and I
 would avoid tying the problem to `exclude` as the same behaviour exists
 for `filter` as well as demonstrated in [https://dbfiddle.uk/i1Kd1GDe this
 fiddle].

 As alluded to ticket:20024#comment:14 fixing this for all nullable
 expressions is quite hard to do the right way. Even if we reduced the
 problem to only queryset like right-hand-side we'd still have to find the
 column being matched against and alter the query to do the equivalent of
 `filter(inner_field__isnull=False)` without causing side effects (e.g.
 what if the field is an aggregation or window function?).
-- 
Ticket URL: <https://code.djangoproject.com/ticket/37140#comment:3>
Django <https://code.djangoproject.com/>
The Web framework for perfectionists with deadlines.

-- 
You received this message because you are subscribed to the Google Groups 
"Django updates" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/django-updates/0107019ec72885a5-0ade914b-ae62-4d44-9117-989026df59c8-000000%40eu-central-1.amazonses.com.

Reply via email to