Fixed #3615: Added support for loading fixtures with forward references on database backends (such as MySQL/InnoDB) that do not support deferred constraint checking. Many thanks to jsdalton for coming up with a clever solution to this long-standing issue, and to jacob, ramiro, graham_king, and russellm for review/testing. (Apologies if I missed anyone else who helped here.)

git-svn-id: http://code.djangoproject.com/svn/django/trunk@16590 bcc190cf-cafb-0310-a4f2-bffc1f526a37

9 files changed, 229 insertions, 16 deletions

diff --git a/django/core/management/commands/loaddata.py b/django/core/management/commands/loaddata.py
index 34f354341e..952b1cb20a 100644
--- a/django/core/management/commands/loaddata.py
+++ b/django/core/management/commands/loaddata.py
@@ -1,3 +1,7 @@
+# This is necessary in Python 2.5 to enable the with statement, in 2.6
+# and up it is no longer necessary.
+from __future__ import with_statement
+
 import sys
 import os
 import gzip
@@ -166,12 +170,20 @@ class Command(BaseCommand):
                                     (format, fixture_name, humanize(fixture_dir)))
                             try:
                                 objects = serializers.deserialize(format, fixture, using=using)
-                                for obj in objects:
-                                    objects_in_fixture += 1
-                                    if router.allow_syncdb(using, obj.object.__class__):
-                                        loaded_objects_in_fixture += 1
-                                        models.add(obj.object.__class__)
-                                        obj.save(using=using)
+
+                                with connection.constraint_checks_disabled():
+                                    for obj in objects:
+                                        objects_in_fixture += 1
+                                        if router.allow_syncdb(using, obj.object.__class__):
+                                            loaded_objects_in_fixture += 1
+                                            models.add(obj.object.__class__)
+                                            obj.save(using=using)
+
+                                # Since we disabled constraint checks, we must manually check for
+                                # any invalid keys that might have been added
+                                table_names = [model._meta.db_table for model in models]
+                                connection.check_constraints(table_names=table_names)
+
                                 loaded_object_count += loaded_objects_in_fixture
                                 fixture_object_count += objects_in_fixture
                                 label_found = True
diff --git a/django/db/backends/__init__.py b/django/db/backends/__init__.py
index 1c3bc7e919..23ddedb4c6 100644
--- a/django/db/backends/__init__.py
+++ b/django/db/backends/__init__.py
@@ -3,6 +3,7 @@ try:
 except ImportError:
     import dummy_thread as thread
 from threading import local
+from contextlib import contextmanager
 
 from django.conf import settings
 from django.db import DEFAULT_DB_ALIAS
@@ -238,6 +239,35 @@ class BaseDatabaseWrapper(local):
         if self.savepoint_state:
             self._savepoint_commit(sid)
 
+    @contextmanager
+    def constraint_checks_disabled(self):
+        disabled = self.disable_constraint_checking()
+        try:
+            yield
+        finally:
+            if disabled:
+                self.enable_constraint_checking()
+
+    def disable_constraint_checking(self):
+        """
+        Backends can implement as needed to temporarily disable foreign key constraint
+        checking.
+        """
+        pass
+
+    def enable_constraint_checking(self):
+        """
+        Backends can implement as needed to re-enable foreign key constraint checking.
+        """
+        pass
+
+    def check_constraints(self, table_names=None):
+        """
+        Backends can override this method if they can apply constraint checking (e.g. via "SET CONSTRAINTS
+        ALL IMMEDIATE"). Should raise an IntegrityError if any invalid foreign key references are encountered.
+        """
+        pass
+
     def close(self):
         if self.connection is not None:
             self.connection.close()
@@ -869,6 +899,19 @@ class BaseDatabaseIntrospection(object):
 
         return sequence_list
 
+    def get_key_columns(self, cursor, table_name):
+        """
+        Backends can override this to return a list of (column_name, referenced_table_name,
+        referenced_column_name) for all key columns in given table.
+        """
+        raise NotImplementedError
+
+    def get_primary_key_column(self, cursor, table_name):
+        """
+        Backends can override this to return the column name of the primary key for the given table.
+        """
+        raise NotImplementedError
+
 class BaseDatabaseClient(object):
     """
     This class encapsulates all backend-specific methods for opening a
diff --git a/django/db/backends/dummy/base.py b/django/db/backends/dummy/base.py
index 7de48c8b00..746f26bacc 100644
--- a/django/db/backends/dummy/base.py
+++ b/django/db/backends/dummy/base.py
@@ -34,6 +34,7 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
     get_table_description = complain
     get_relations = complain
     get_indexes = complain
+    get_key_columns = complain
 
 class DatabaseWrapper(BaseDatabaseWrapper):
     operators = {}
diff --git a/django/db/backends/mysql/base.py b/django/db/backends/mysql/base.py
index 6d02aa771c..f4523e4e82 100644
--- a/django/db/backends/mysql/base.py
+++ b/django/db/backends/mysql/base.py
@@ -349,3 +349,52 @@ class DatabaseWrapper(BaseDatabaseWrapper):
                 raise Exception('Unable to determine MySQL version from version string %r' % self.connection.get_server_info())
             self.server_version = tuple([int(x) for x in m.groups()])
         return self.server_version
+
+    def disable_constraint_checking(self):
+        """
+        Disables foreign key checks, primarily for use in adding rows with forward references. Always returns True,
+        to indicate constraint checks need to be re-enabled.
+        """
+        self.cursor().execute('SET foreign_key_checks=0')
+        return True
+
+    def enable_constraint_checking(self):
+        """
+        Re-enable foreign key checks after they have been disabled.
+        """
+        self.cursor().execute('SET foreign_key_checks=1')
+
+    def check_constraints(self, table_names=None):
+        """
+        Checks each table name in table-names for rows with invalid foreign key references. This method is
+        intended to be used in conjunction with `disable_constraint_checking()` and `enable_constraint_checking()`, to
+        determine if rows with invalid references were entered while constraint checks were off.
+
+        Raises an IntegrityError on the first invalid foreign key reference encountered (if any) and provides
+        detailed information about the invalid reference in the error message.
+
+        Backends can override this method if they can more directly apply constraint checking (e.g. via "SET CONSTRAINTS
+        ALL IMMEDIATE")
+        """
+        cursor = self.cursor()
+        if table_names is None:
+            table_names = self.introspection.get_table_list(cursor)
+        for table_name in table_names:
+            primary_key_column_name = self.introspection.get_primary_key_column(cursor, table_name)
+            if not primary_key_column_name:
+                continue
+            key_columns = self.introspection.get_key_columns(cursor, table_name)
+            for column_name, referenced_table_name, referenced_column_name in key_columns:
+                cursor.execute("""
+                    SELECT REFERRING.`%s`, REFERRING.`%s` FROM `%s` as REFERRING
+                    LEFT JOIN `%s` as REFERRED
+                    ON (REFERRING.`%s` = REFERRED.`%s`)
+                    WHERE REFERRING.`%s` IS NOT NULL AND REFERRED.`%s` IS NULL"""
+                    % (primary_key_column_name, column_name, table_name, referenced_table_name,
+                    column_name, referenced_column_name, column_name, referenced_column_name))
+                for bad_row in cursor.fetchall():
+                    raise utils.IntegrityError("The row in table '%s' with primary key '%s' has an invalid "
+                        "foreign key: %s.%s contains a value '%s' that does not have a corresponding value in %s.%s."
+                        % (table_name, bad_row[0],
+                        table_name, column_name, bad_row[1],
+                        referenced_table_name, referenced_column_name))
diff --git a/django/db/backends/mysql/introspection.py b/django/db/backends/mysql/introspection.py
index 9e1518b06e..ab4eebea90 100644
--- a/django/db/backends/mysql/introspection.py
+++ b/django/db/backends/mysql/introspection.py
@@ -51,10 +51,21 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
         representing all relationships to the given table. Indexes are 0-based.
         """
         my_field_dict = self._name_to_index(cursor, table_name)
-        constraints = []
+        constraints = self.get_key_columns(cursor, table_name)
         relations = {}
+        for my_fieldname, other_table, other_field in constraints:
+            other_field_index = self._name_to_index(cursor, other_table)[other_field]
+            my_field_index = my_field_dict[my_fieldname]
+            relations[my_field_index] = (other_field_index, other_table)
+        return relations
+
+    def get_key_columns(self, cursor, table_name):
+        """
+        Returns a list of (column_name, referenced_table_name, referenced_column_name) for all
+        key columns in given table.
+        """
+        key_columns = []
         try:
-            # This should work for MySQL 5.0.
             cursor.execute("""
                 SELECT column_name, referenced_table_name, referenced_column_name
                 FROM information_schema.key_column_usage
@@ -62,7 +73,7 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
                     AND table_schema = DATABASE()
                     AND referenced_table_name IS NOT NULL
                     AND referenced_column_name IS NOT NULL""", [table_name])
-            constraints.extend(cursor.fetchall())
+            key_columns.extend(cursor.fetchall())
         except (ProgrammingError, OperationalError):
             # Fall back to "SHOW CREATE TABLE", for previous MySQL versions.
             # Go through all constraints and save the equal matches.
@@ -74,14 +85,17 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
                     if match == None:
                         break
                     pos = match.end()
-                    constraints.append(match.groups())
+                    key_columns.append(match.groups())
+        return key_columns
 
-        for my_fieldname, other_table, other_field in constraints:
-            other_field_index = self._name_to_index(cursor, other_table)[other_field]
-            my_field_index = my_field_dict[my_fieldname]
-            relations[my_field_index] = (other_field_index, other_table)
-
-        return relations
+    def get_primary_key_column(self, cursor, table_name):
+        """
+        Returns the name of the primary key column for the given table
+        """
+        for column in self.get_indexes(cursor, table_name).iteritems():
+            if column[1]['primary_key']:
+                return column[0]
+        return None
 
     def get_indexes(self, cursor, table_name):
         """
diff --git a/django/db/backends/oracle/base.py b/django/db/backends/oracle/base.py
index 930b1bb0f1..3cadb6617d 100644
--- a/django/db/backends/oracle/base.py
+++ b/django/db/backends/oracle/base.py
@@ -428,6 +428,14 @@ class DatabaseWrapper(BaseDatabaseWrapper):
         self.introspection = DatabaseIntrospection(self)
         self.validation = BaseDatabaseValidation(self)
 
+    def check_constraints(self, table_names=None):
+        """
+        To check constraints, we set constraints to immediate. Then, when, we're done we must ensure they
+        are returned to deferred.
+        """
+        self.cursor().execute('SET CONSTRAINTS ALL IMMEDIATE')
+        self.cursor().execute('SET CONSTRAINTS ALL DEFERRED')
+
     def _valid_connection(self):
         return self.connection is not None
 
diff --git a/django/db/backends/postgresql_psycopg2/base.py b/django/db/backends/postgresql_psycopg2/base.py
index 6ed59a66f4..37aa072f6b 100644
--- a/django/db/backends/postgresql_psycopg2/base.py
+++ b/django/db/backends/postgresql_psycopg2/base.py
@@ -106,6 +106,14 @@ class DatabaseWrapper(BaseDatabaseWrapper):
         self.validation = BaseDatabaseValidation(self)
         self._pg_version = None
 
+    def check_constraints(self, table_names=None):
+        """
+        To check constraints, we set constraints to immediate. Then, when, we're done we must ensure they
+        are returned to deferred.
+        """
+        self.cursor().execute('SET CONSTRAINTS ALL IMMEDIATE')
+        self.cursor().execute('SET CONSTRAINTS ALL DEFERRED')
+
     def _get_pg_version(self):
         if self._pg_version is None:
             self._pg_version = get_version(self.connection)
diff --git a/django/db/backends/sqlite3/base.py b/django/db/backends/sqlite3/base.py
index 79c5eded34..f922d38638 100644
--- a/django/db/backends/sqlite3/base.py
+++ b/django/db/backends/sqlite3/base.py
@@ -206,6 +206,40 @@ class DatabaseWrapper(BaseDatabaseWrapper):
             connection_created.send(sender=self.__class__, connection=self)
         return self.connection.cursor(factory=SQLiteCursorWrapper)
 
+    def check_constraints(self, table_names=None):
+        """
+        Checks each table name in table-names for rows with invalid foreign key references. This method is
+        intended to be used in conjunction with `disable_constraint_checking()` and `enable_constraint_checking()`, to
+        determine if rows with invalid references were entered while constraint checks were off.
+
+        Raises an IntegrityError on the first invalid foreign key reference encountered (if any) and provides
+        detailed information about the invalid reference in the error message.
+
+        Backends can override this method if they can more directly apply constraint checking (e.g. via "SET CONSTRAINTS
+        ALL IMMEDIATE")
+        """
+        cursor = self.cursor()
+        if table_names is None:
+            table_names = self.introspection.get_table_list(cursor)
+        for table_name in table_names:
+            primary_key_column_name = self.introspection.get_primary_key_column(cursor, table_name)
+            if not primary_key_column_name:
+                continue
+            key_columns = self.introspection.get_key_columns(cursor, table_name)
+            for column_name, referenced_table_name, referenced_column_name in key_columns:
+                cursor.execute("""
+                    SELECT REFERRING.`%s`, REFERRING.`%s` FROM `%s` as REFERRING
+                    LEFT JOIN `%s` as REFERRED
+                    ON (REFERRING.`%s` = REFERRED.`%s`)
+                    WHERE REFERRING.`%s` IS NOT NULL AND REFERRED.`%s` IS NULL"""
+                    % (primary_key_column_name, column_name, table_name, referenced_table_name,
+                    column_name, referenced_column_name, column_name, referenced_column_name))
+                for bad_row in cursor.fetchall():
+                    raise utils.IntegrityError("The row in table '%s' with primary key '%s' has an invalid "
+                        "foreign key: %s.%s contains a value '%s' that does not have a corresponding value in %s.%s."
+                        % (table_name, bad_row[0], table_name, column_name, bad_row[1],
+                        referenced_table_name, referenced_column_name))
+
     def close(self):
         # If database is in memory, closing the connection destroys the
         # database. To prevent accidental data loss, ignore close requests on
diff --git a/django/db/backends/sqlite3/introspection.py b/django/db/backends/sqlite3/introspection.py
index 5ee7b64bcd..9652a4da6a 100644
--- a/django/db/backends/sqlite3/introspection.py
+++ b/django/db/backends/sqlite3/introspection.py
@@ -103,6 +103,35 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
 
         return relations
 
+    def get_key_columns(self, cursor, table_name):
+        """
+        Returns a list of (column_name, referenced_table_name, referenced_column_name) for all
+        key columns in given table.
+        """
+        key_columns = []
+
+        # Schema for this table
+        cursor.execute("SELECT sql FROM sqlite_master WHERE tbl_name = %s AND type = %s", [table_name, "table"])
+        results = cursor.fetchone()[0].strip()
+        results = results[results.index('(')+1:results.rindex(')')]
+
+        # Walk through and look for references to other tables. SQLite doesn't
+        # really have enforced references, but since it echoes out the SQL used
+        # to create the table we can look for REFERENCES statements used there.
+        for field_index, field_desc in enumerate(results.split(',')):
+            field_desc = field_desc.strip()
+            if field_desc.startswith("UNIQUE"):
+                continue
+
+            m = re.search('"(.*)".*references (.*) \(["|](.*)["|]\)', field_desc, re.I)
+            if not m:
+                continue
+
+            # This will append (column_name, referenced_table_name, referenced_column_name) to key_columns
+            key_columns.append(tuple([s.strip('"') for s in m.groups()]))
+
+        return key_columns
+
     def get_indexes(self, cursor, table_name):
         """
         Returns a dictionary of fieldname -> infodict for the given table,
@@ -128,6 +157,21 @@ class DatabaseIntrospection(BaseDatabaseIntrospection):
             indexes[name]['unique'] = True
         return indexes
 
+    def get_primary_key_column(self, cursor, table_name):
+        """
+        Get the column name of the primary key for the given table.
+        """
+        # Don't use PRAGMA because that causes issues with some transactions
+        cursor.execute("SELECT sql FROM sqlite_master WHERE tbl_name = %s AND type = %s", [table_name, "table"])
+        results = cursor.fetchone()[0].strip()
+        results = results[results.index('(')+1:results.rindex(')')]
+        for field_desc in results.split(','):
+            field_desc = field_desc.strip()
+            m = re.search('"(.*)".*PRIMARY KEY$', field_desc)
+            if m:
+                return m.groups()[0]
+        return None
+
     def _table_info(self, cursor, name):
         cursor.execute('PRAGMA table_info(%s)' % self.connection.ops.quote_name(name))
         # cid, name, type, notnull, dflt_value, pk