Fixed #14244: Allow lists of more than 1000 items to be used with the 'in' lookup in Oracle, by breaking them up into groups of 1000 items and ORing them together. Thanks to rlynch for the report and initial patch.

git-svn-id: http://code.djangoproject.com/svn/django/trunk@13859 bcc190cf-cafb-0310-a4f2-bffc1f526a37

django/db/backends/__init__.py

@@ -233,6 +233,13 @@ class BaseDatabaseOperations(object):
         """
         return "%s"
 
+    def max_in_list_size(self):
+        """
+        Returns the maximum number of items that can be passed in a single 'IN'
+        list condition, or None if the backend does not impose a limit.
+        """
+        return None
+
     def max_name_length(self):
         """
         Returns the maximum length of table and column names, or None if there

django/db/backends/oracle/base.py

@@ -178,6 +178,9 @@ WHEN (new.%(col_name)s IS NULL)
             return "UPPER(%s)"
         return "%s"
 
+    def max_in_list_size(self):
+        return 1000
+
     def max_name_length(self):
         return 30
 

django/db/models/sql/where.py

@@ -2,6 +2,7 @@
 Code to manage the creation and SQL rendering of 'where' constraints.
 """
 import datetime
+from itertools import repeat
 
 from django.utils import tree
 from django.db.models.fields import Field
@@ -178,8 +179,24 @@ class WhereNode(tree.Node):
                 raise EmptyResultSet
             if extra:
                 return ('%s IN %s' % (field_sql, extra), params)
-            return ('%s IN (%s)' % (field_sql, ', '.join(['%s'] * len(params))),
-                    params)
+            max_in_list_size = connection.ops.max_in_list_size()
+            if max_in_list_size and len(params) > max_in_list_size:
+                # Break up the params list into an OR of manageable chunks.
+                in_clause_elements = ['(']
+                for offset in xrange(0, len(params), max_in_list_size):
+                    if offset > 0:
+                        in_clause_elements.append(' OR ')
+                    in_clause_elements.append('%s IN (' % field_sql)
+                    group_size = min(len(params) - offset, max_in_list_size)
+                    param_group = ', '.join(repeat('%s', group_size))
+                    in_clause_elements.append(param_group)
+                    in_clause_elements.append(')')
+                in_clause_elements.append(')')
+                return ''.join(in_clause_elements), params
+            else:
+                return ('%s IN (%s)' % (field_sql,
+                                        ', '.join(repeat('%s', len(params)))),
+                        params)
         elif lookup_type in ('range', 'year'):
             return ('%s BETWEEN %%s and %%s' % field_sql, params)
         elif lookup_type in ('month', 'day', 'week_day'):

tests/regressiontests/queries/models.py

@@ -1339,3 +1339,23 @@ Using an empty generator expression as the rvalue for an "__in" lookup is legal
 []
 
 """
+
+# Sqlite 3 does not support passing in more than 1000 parameters except by
+# changing a parameter at compilation time.
+if settings.DATABASES[DEFAULT_DB_ALIAS]['ENGINE'] != "django.db.backends.sqlite3":
+    __test__["API_TESTS"] += """
+Bug #14244: Test that the "in" lookup works with lists of 1000 items or more.
+>>> Number.objects.all().delete()
+>>> numbers = range(2500)
+>>> for num in numbers:
+...     _ = Number.objects.create(num=num)
+>>> Number.objects.filter(num__in=numbers[:1000]).count()
+1000
+>>> Number.objects.filter(num__in=numbers[:1001]).count()
+1001
+>>> Number.objects.filter(num__in=numbers[:2000]).count()
+2000
+>>> Number.objects.filter(num__in=numbers).count()
+2500
+
+"""