Fixed #19500 -- Solved a regression in join reuse

The ORM didn't reuse joins for direct foreign key traversals when using
chained filters. For example:
    qs.filter(fk__somefield=1).filter(fk__somefield=2))
produced two joins.

As a bonus, reverse onetoone filters can now reuse joins correctly

The regression was caused by the join() method refactor in commit
68847135bc

Thanks for Simon Charette for spotting some issues with the first draft
of the patch.

django/db/models/sql/compiler.py

@@ -6,7 +6,7 @@ from django.db.backends.util import truncate_name
 from django.db.models.constants import LOOKUP_SEP
 from django.db.models.query_utils import select_related_descend
 from django.db.models.sql.constants import (SINGLE, MULTI, ORDER_DIR,
-        GET_ITERATOR_CHUNK_SIZE, REUSE_ALL, SelectInfo)
+        GET_ITERATOR_CHUNK_SIZE, SelectInfo)
 from django.db.models.sql.datastructures import EmptyResultSet
 from django.db.models.sql.expressions import SQLEvaluator
 from django.db.models.sql.query import get_order_dir, Query
@@ -317,7 +317,7 @@ class SQLCompiler(object):
 
         for name in self.query.distinct_fields:
             parts = name.split(LOOKUP_SEP)
-            field, col, alias, _, _ = self._setup_joins(parts, opts, None)
+            field, col, alias, _, _ = self._setup_joins(parts, opts)
             col, alias = self._final_join_removal(col, alias)
             result.append("%s.%s" % (qn(alias), qn2(col)))
         return result
@@ -450,7 +450,7 @@ class SQLCompiler(object):
         if not alias:
             alias = self.query.get_initial_alias()
         field, target, opts, joins, _ = self.query.setup_joins(
-            pieces, opts, alias, REUSE_ALL)
+            pieces, opts, alias)
         # We will later on need to promote those joins that were added to the
         # query afresh above.
         joins_to_promote = [j for j in joins if self.query.alias_refcount[j] < 2]
@@ -688,7 +688,7 @@ class SQLCompiler(object):
                         int_opts = int_model._meta
                         alias = self.query.join(
                             (alias, int_opts.db_table, lhs_col, int_opts.pk.column),
-                            promote=True,
+                            promote=True
                         )
                         alias_chain.append(alias)
                 alias = self.query.join(

django/db/models/sql/constants.py

@@ -44,6 +44,3 @@ ORDER_DIR = {
     'ASC': ('ASC', 'DESC'),
     'DESC': ('DESC', 'ASC'),
 }
-
-# A marker for join-reusability.
-REUSE_ALL = object()

django/db/models/sql/expressions.py

@@ -1,10 +1,9 @@
 from django.core.exceptions import FieldError
 from django.db.models.constants import LOOKUP_SEP
 from django.db.models.fields import FieldDoesNotExist
-from django.db.models.sql.constants import REUSE_ALL
 
 class SQLEvaluator(object):
-    def __init__(self, expression, query, allow_joins=True, reuse=REUSE_ALL):
+    def __init__(self, expression, query, allow_joins=True, reuse=None):
         self.expression = expression
         self.opts = query.get_meta()
         self.cols = []
@@ -54,7 +53,7 @@ class SQLEvaluator(object):
                     field_list, query.get_meta(),
                     query.get_initial_alias(), self.reuse)
                 col, _, join_list = query.trim_joins(source, join_list, path)
-                if self.reuse is not None and self.reuse != REUSE_ALL:
+                if self.reuse is not None:
                     self.reuse.update(join_list)
                 self.cols.append((node, (join_list[-1], col)))
             except FieldDoesNotExist:

django/db/models/sql/query.py

@@ -20,7 +20,7 @@ from django.db.models.fields import FieldDoesNotExist
 from django.db.models.loading import get_model
 from django.db.models.sql import aggregates as base_aggregates_module
 from django.db.models.sql.constants import (QUERY_TERMS, ORDER_DIR, SINGLE,
-        ORDER_PATTERN, REUSE_ALL, JoinInfo, SelectInfo, PathInfo)
+        ORDER_PATTERN, JoinInfo, SelectInfo, PathInfo)
 from django.db.models.sql.datastructures import EmptyResultSet, Empty, MultiJoin
 from django.db.models.sql.expressions import SQLEvaluator
 from django.db.models.sql.where import (WhereNode, Constraint, EverythingNode,
@@ -891,7 +891,7 @@ class Query(object):
         """
         return len([1 for count in self.alias_refcount.values() if count])
 
-    def join(self, connection, reuse=REUSE_ALL, promote=False,
+    def join(self, connection, reuse=None, promote=False,
              outer_if_first=False, nullable=False, join_field=None):
         """
         Returns an alias for the join in 'connection', either reusing an
@@ -902,10 +902,9 @@ class Query(object):
 
             lhs.lhs_col = table.col
 
-        The 'reuse' parameter can be used in three ways: it can be REUSE_ALL
-        which means all joins (matching the connection) are reusable, it can
-        be a set containing the aliases that can be reused, or it can be None
-        which means a new join is always created.
+        The 'reuse' parameter can be either None which means all joins
+        (matching the connection) are reusable, or it can be a set containing
+        the aliases that can be reused.
 
         If 'promote' is True, the join type for the alias will be LOUTER (if
         the alias previously existed, the join type will be promoted from INNER
@@ -926,10 +925,8 @@ class Query(object):
         """
         lhs, table, lhs_col, col = connection
         existing = self.join_map.get(connection, ())
-        if reuse == REUSE_ALL:
+        if reuse is None:
             reuse = existing
-        elif reuse is None:
-            reuse = set()
         else:
             reuse = [a for a in existing if a in reuse]
         for alias in reuse:
@@ -1040,7 +1037,7 @@ class Query(object):
             # then we need to explore the joins that are required.
 
             field, source, opts, join_list, path = self.setup_joins(
-                field_list, opts, self.get_initial_alias(), REUSE_ALL)
+                field_list, opts, self.get_initial_alias())
 
             # Process the join chain to see if it can be trimmed
             col, _, join_list = self.trim_joins(source, join_list, path)
@@ -1441,7 +1438,7 @@ class Query(object):
             raise MultiJoin(multijoin_pos + 1)
         return path, final_field, target
 
-    def setup_joins(self, names, opts, alias, can_reuse, allow_many=True,
+    def setup_joins(self, names, opts, alias, can_reuse=None, allow_many=True,
                     allow_explicit_fk=False):
         """
         Compute the necessary table joins for the passage through the fields
@@ -1450,9 +1447,9 @@ class Query(object):
         the table to start the joining from.
 
         The 'can_reuse' defines the reverse foreign key joins we can reuse. It
-        can be sql.constants.REUSE_ALL in which case all joins are reusable
-        or a set of aliases that can be reused. Note that Non-reverse foreign
-        keys are always reusable.
+        can be None in which case all joins are reusable or a set of aliases
+        that can be reused. Note that non-reverse foreign keys are always
+        reusable when using setup_joins().
 
         If 'allow_many' is False, then any reverse foreign key seen will
         generate a MultiJoin exception.
@@ -1485,8 +1482,9 @@ class Query(object):
             else:
                 nullable = True
             connection = alias, opts.db_table, from_field.column, to_field.column
-            alias = self.join(connection, reuse=can_reuse, nullable=nullable,
-                              join_field=join_field)
+            reuse = None if direct or to_field.unique else can_reuse
+            alias = self.join(connection, reuse=reuse,
+                              nullable=nullable, join_field=join_field)
             joins.append(alias)
         return final_field, target, opts, joins, path
 
@@ -1643,7 +1641,7 @@ class Query(object):
         try:
             for name in field_names:
                 field, target, u2, joins, u3 = self.setup_joins(
-                        name.split(LOOKUP_SEP), opts, alias, REUSE_ALL, allow_m2m,
+                        name.split(LOOKUP_SEP), opts, alias, None, allow_m2m,
                         True)
                 final_alias = joins[-1]
                 col = target.column
@@ -1729,8 +1727,9 @@ class Query(object):
         else:
             opts = self.model._meta
             if not self.select:
-                count = self.aggregates_module.Count((self.join((None, opts.db_table, None, None)), opts.pk.column),
-                                         is_summary=True, distinct=True)
+                count = self.aggregates_module.Count(
+                    (self.join((None, opts.db_table, None, None)), opts.pk.column),
+                    is_summary=True, distinct=True)
             else:
                 # Because of SQL portability issues, multi-column, distinct
                 # counts need a sub-query -- see get_count() for details.
@@ -1934,7 +1933,7 @@ class Query(object):
         opts = self.model._meta
         alias = self.get_initial_alias()
         field, col, opts, joins, extra = self.setup_joins(
-                start.split(LOOKUP_SEP), opts, alias, REUSE_ALL)
+                start.split(LOOKUP_SEP), opts, alias)
         select_col = self.alias_map[joins[1]].lhs_join_col
         select_alias = alias
 

django/db/models/sql/subqueries.py

@@ -232,7 +232,6 @@ class DateQuery(Query):
                 field_name.split(LOOKUP_SEP),
                 self.get_meta(),
                 self.get_initial_alias(),
-                False
             )
         except FieldError:
             raise FieldDoesNotExist("%s has no field named '%s'" % (

tests/regressiontests/generic_relations_regress/tests.py

@@ -72,5 +72,11 @@ class GenericRelationTests(TestCase):
             Q(notes__note__icontains=r'other note'))
         self.assertTrue(org_contact in qs)
 
-
+    def test_join_reuse(self):
+        qs = Person.objects.filter(
+            addresses__street='foo'
+        ).filter(
+            addresses__street='bar'
+        )
+        self.assertEqual(str(qs.query).count('JOIN'), 2)
 

tests/regressiontests/queries/tests.py

@@ -2418,3 +2418,36 @@ class ReverseJoinTrimmingTest(TestCase):
         qs = Tag.objects.filter(annotation__tag=t.pk)
         self.assertIn('INNER JOIN', str(qs.query))
         self.assertEquals(list(qs), [])
+
+class JoinReuseTest(TestCase):
+    """
+    Test that the queries reuse joins sensibly (for example, direct joins
+    are always reused).
+    """
+    def test_fk_reuse(self):
+        qs = Annotation.objects.filter(tag__name='foo').filter(tag__name='bar')
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_fk_reuse_select_related(self):
+        qs = Annotation.objects.filter(tag__name='foo').select_related('tag')
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_fk_reuse_annotation(self):
+        qs = Annotation.objects.filter(tag__name='foo').annotate(cnt=Count('tag__name'))
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_fk_reuse_disjunction(self):
+        qs = Annotation.objects.filter(Q(tag__name='foo') | Q(tag__name='bar'))
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_fk_reuse_order_by(self):
+        qs = Annotation.objects.filter(tag__name='foo').order_by('tag__name')
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_revo2o_reuse(self):
+        qs = Detail.objects.filter(member__name='foo').filter(member__name='foo')
+        self.assertEqual(str(qs.query).count('JOIN'), 1)
+
+    def test_revfk_noreuse(self):
+        qs = Author.objects.filter(report__name='r4').filter(report__name='r1')
+        self.assertEqual(str(qs.query).count('JOIN'), 2)