HAVING filter screws with extra() SQL parameter ordering

[miracle2k]

>>> qs = Model.objects.extra(where=['foo>%s'], params=[2]).annotate(count=Count('bar')).filter(bar=99)
>>> qs.query.as_sql()
('SELECT `site_publication`.`id`, ..., FROM ... WHERE foo > %s GROUP BY ... HAVING bar > %s',
 (99, 2))

As you can see, the parameters (99 and 2) are in the wrong order. "foo", the first filter, should be evaluated against 2, and bar, the second one, against 99.

If "bar" is not an aggregate, it works, because apparently the filter would be placed in the WHERE clause *before* the "extra()" where.

Using rev. 10743.

[Russell Keith-Magee]

A related use case, reported in #11117:

Book.objects.extra(
    select={'date': "DATE_FORMAT(timestamp, '%s')"}, 
    select_params=('%%H',), 
    where=["DATE_FORMAT(timestamp, '%%H:%%i') > '%s'"],
    params=['22:00']).values('date').annotate(cnt=Count('id'))

the resulting sql query is:

SELECT DATE_FORMAT(timestamp, '%H') AS `date`, COUNT(id) AS `cnt` 
FROM `books`
WHERE DATE_FORMAT(timestamp, '%H:%i') > '%H' 
GROUP BY DATE_FORMAT(timestamp, '22:00') 
ORDER BY timestamp ASC

The placeholder values '%H' and '22:00' are interchanged, which is obviously a result of using the GROUP BY here on a user-defined variable with placeholder (--> 'date')

[Marco Paolini]

I think all filters on annotated queryset should end up in HAVING clause, not in WHERE

imagine you want to do something like:

SELECT * FROM blog group by location, rating having count(id) > 10 OR rating >1 

right now (r11199) you can ony add HAVING clauses using if you filter using an aggregate. Does this make sense to you?

[Alex Gaynor]

#11916 was closed as a dupe of this, it has a patch.

[paluh]

changing grouping generation from values to keys of extra selects

[paluh]

I think that I've found bug related to this ticket. When generating list for grouping values (not keys) are taken from extra selects.

--- django/db/models/sql/query.py	(revision 11729)
+++ django/db/models/sql/query.py	(working copy)
@@ -885,9 +885,9 @@
             group_by = self.group_by or []
 
             extra_selects = []
-            for extra_select, extra_params in self.extra_select.itervalues():
-                extra_selects.append(extra_select)
-                params.extend(extra_params)
+            for extra_select_key in self.extra_select.iterkeys():
+                extra_selects.append(extra_select_key)
+
             for col in group_by + self.related_select_cols + extra_selects:
                 if isinstance(col, (list, tuple)):
                     result.append('%s.%s' % (qn(col[0]), qn(col[1])))

[paluh]

sorry for my previous post and attachment - it's related to #11916 which I've reopened.

[Aymeric Augustin]

Change UI/UX from NULL to False.

[Aymeric Augustin]

Change Easy pickings from NULL to False.

[Tim Graham]

We are no longer fixing bugs with QuerySet.extra() per ​discussion on django-developers.