Add regex support

Author: OskarPersson

sql_server/pyodbc/creation.py

@@ -1,3 +1,6 @@
+import binascii
+import os
+
 from django.db.backends.base.creation import BaseDatabaseCreation
 
 
@@ -25,3 +28,68 @@ def sql_table_creation_suffix(self):
         if collation:
             suffix.append('COLLATE %s' % collation)
         return ' '.join(suffix)
+
+    # The following code to add regex support in SQLServer is taken from django-mssql
+    # see https://bitbucket.org/Manfre/django-mssql
+    def enable_clr(self):
+        """ Enables clr for server if not already enabled
+        This function will not fail if current user doesn't have
+        permissions to enable clr, and clr is already enabled
+        """
+        with self._nodb_connection.cursor() as cursor:
+            # check whether clr is enabled
+            cursor.execute('''
+            SELECT value FROM sys.configurations
+            WHERE name = 'clr enabled'
+            ''')
+            res = None
+            try:
+                res = cursor.fetchone()
+            except Exception:
+                pass
+
+            if not res or not res[0]:
+                # if not enabled enable clr
+                cursor.execute("sp_configure 'clr enabled', 1")
+                cursor.execute("RECONFIGURE")
+
+                cursor.execute("sp_configure 'show advanced options', 1")
+                cursor.execute("RECONFIGURE")
+
+                cursor.execute("sp_configure 'clr strict security', 0")
+                cursor.execute("RECONFIGURE")
+
+    def install_regex_clr(self, database_name):
+        sql = '''
+USE {database_name};
+-- Drop and recreate the function if it already exists
+IF OBJECT_ID('REGEXP_LIKE') IS NOT NULL
+DROP FUNCTION [dbo].[REGEXP_LIKE]
+IF EXISTS(select * from sys.assemblies where name like 'regex_clr')
+DROP ASSEMBLY regex_clr
+;
+CREATE ASSEMBLY regex_clr
+FROM 0x{assembly_hex}
+WITH PERMISSION_SET = SAFE;
+create function [dbo].[REGEXP_LIKE]
+(
+@input nvarchar(max),
+@pattern nvarchar(max),
+@caseSensitive int
+)
+RETURNS INT  AS
+EXTERNAL NAME regex_clr.UserDefinedFunctions.REGEXP_LIKE
+        '''.format(
+            database_name=self.connection.ops.quote_name(database_name),
+            assembly_hex=self.get_regex_clr_assembly_hex(),
+        ).split(';')
+
+        self.enable_clr()
+
+        with self._nodb_connection.cursor() as cursor:
+            for s in sql:
+                cursor.execute(s)
+
+    def get_regex_clr_assembly_hex(self):
+        with open(os.path.join(os.path.dirname(__file__), 'regex_clr.dll'), 'rb') as f:
+            return binascii.hexlify(f.read()).decode('ascii')

sql_server/pyodbc/features.py

@@ -25,7 +25,7 @@ class DatabaseFeatures(BaseDatabaseFeatures):
     supports_ignore_conflicts = False
     supports_index_on_text_field = False
     supports_paramstyle_pyformat = False
-    supports_regex_backreferencing = False
+    supports_regex_backreferencing = True
     supports_sequence_reset = False
     supports_subqueries_in_group_by = False
     supports_tablespaces = True

sql_server/pyodbc/management/__init__.py

@@ -0,0 +1,25 @@
+# Add regex support in SQLServer
+# Code taken from django-mssql (see https://bitbucket.org/Manfre/django-mssql)
+
+from django.core.management.base import BaseCommand
+from django.db import connection
+
+
+class Command(BaseCommand):
+    help = "Installs the regex_clr.dll assembly with the database"
+
+    requires_model_validation = False
+
+    args = 'database_name'
+
+    def add_arguments(self, parser):
+        parser.add_argument('database_name')
+
+    def handle(self, *args, **options):
+        database_name = options['database_name']
+        if not database_name:
+            self.print_help('manage.py', 'install_regex_clr')
+            return
+
+        connection.creation.install_regex_clr(database_name)
+        print('Installed regex_clr to database %s' % database_name)

sql_server/pyodbc/management/commands/__init__.py

@@ -273,7 +273,8 @@ def regex_lookup(self, lookup_type):
         If the feature is not supported (or part of it is not supported), a
         NotImplementedError exception can be raised.
         """
-        raise NotImplementedError('SQL Server has no built-in regular expression support.')
+        match_option = {'iregex': 0, 'regex': 1}[lookup_type]
+        return "dbo.REGEXP_LIKE(%%s, %%s, %s)=1" % (match_option,)
 
     def limit_offset_sql(self, low_mark, high_mark):
         """Return LIMIT/OFFSET SQL clause."""

sql_server/pyodbc/management/commands/install_regex_clr.py

@@ -12,7 +12,7 @@ git fetch --depth=1 origin +refs/tags/*:refs/tags/*
 git checkout $DJANGO_VERSION
 pip install -r tests/requirements/py3.txt
 
-python tests/runtests.py --settings=testapp.settings --noinput \
+python tests/runtests.py --settings=testapp.settings --noinput --keepdb \
     aggregation \
     aggregation_regress \
     annotations \

sql_server/pyodbc/operations.py

@@ -9,6 +9,7 @@
     'django.contrib.contenttypes',
     'django.contrib.staticfiles',
     'django.contrib.auth',
+    'sql_server.pyodbc',
     'testapp',
 )
 

sql_server/pyodbc/regex_clr.dll

@@ -12,7 +12,8 @@ whitelist_externals =
     /bin/bash
 
 commands =
-    python manage.py test
+    python manage.py test --keepdb
+    python manage.py install_regex_clr test_default
     bash test.sh
 
 deps =