remove counting rows in checking for resync

transferwise · Jul 25, 2024 · 0843b33 · 0843b33
1 parent f099e7f
commit 0843b33
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 19 deletions.
diff --git a/pipelinewise/utils.py b/pipelinewise/utils.py
@@ -30,24 +30,19 @@ def get_tables_size(schema: str, tap) -> dict:
         tap.open_connections()
         result_list = tap.query(
             'select TABLE_NAME as table_name,'
-            ' TABLE_ROWS as table_rows,'
             ' (DATA_LENGTH + INDEX_LENGTH)/ 1024 / 1024 as table_size'
             f' from information_schema.TABLES where TABLE_SCHEMA = \'{schema}\';')
         tap.close_connections()
         for res in result_list:
             result.append({
                 'table_name': f'{schema}.{res["table_name"]}',
-                'table_rows': res['table_rows'],
                 'table_size': res['table_size']
             })
 
     if 'FastSyncTapPostgres' in str(type(tap)):
         tap.open_connection()
         result_list = tap.query(
             'SELECT TABLE_NAME as table_name,'
-            ' (xpath(\'/row/c/text()\','
-            ' query_to_xml(format(\'select count(*) as c from %I.%I\', table_schema, TABLE_NAME), FALSE, TRUE, \'\'))'
-            ')[1]::text::int AS table_rows,'
             ' pg_total_relation_size('
             '\'"\'||table_schema||\'"."\'||table_name||\'"\')::NUMERIC/1024::NUMERIC/1024 as table_size '
             'FROM (SELECT table_schema, TABLE_NAME FROM information_schema.tables '
@@ -56,8 +51,7 @@ def get_tables_size(schema: str, tap) -> dict:
         for res in result_list:
             result.append({
                 'table_name': f'{schema}.{res[0]}',
-                'table_rows': res[1],
-                'table_size': res[2]
+                'table_size': res[1]
             })
         tap.close_connection()
     return result

diff --git a/tests/units/test_utils.py b/tests/units/test_utils.py
@@ -25,14 +25,14 @@ def fetchall(self):
         """Mock fetchall method"""
         if self.type == 'mysql':
             return [
-                {'table_name': 't1', 'table_rows': 1, 'table_size': 1234},
-                {'table_name': 't2', 'table_rows': 3, 'table_size': 1234}
+                {'table_name': 't1', 'table_size': 1234},
+                {'table_name': 't2', 'table_size': 1234}
             ]
 
         if self.type == 'postgres':
             return [
-                ['t1', 1, 1234],
-                ['t2', 3, 1234],
+                ['t1', 1234],
+                ['t2', 1234],
             ]
         return None
 
@@ -119,8 +119,8 @@ def test_get_tables_size(self):
         """Test get_tables_size method works correctly"""
         test_schema = 'foo_schema'
         expected_tables_info = [
-            {'table_name': f'{test_schema}.t1', 'table_rows': 1, 'table_size': 1234},
-            {'table_name': f'{test_schema}.t2', 'table_rows': 3, 'table_size': 1234}
+            {'table_name': f'{test_schema}.t1', 'table_size': 1234},
+            {'table_name': f'{test_schema}.t2', 'table_size': 1234}
         ]
 
         connection_config = {'host': 'foo_host',
@@ -158,15 +158,15 @@ def test_filter_out_selected_tables(self):
         """Test filter_out_selected_tables method works correctly"""
         selected_tables = {'foo.t1', 'foo.t3', 'bar.t2'}
         all_schema_tables = [
-            {'table_name': 'foo.t1', 'table_rows': 1, 'table_size': 1111},
-            {'table_name': 'foo.t2', 'table_rows': 2, 'table_size': 2222},
-            {'table_name': 'foo.t3', 'table_rows': 3, 'table_size': 3333},
-            {'table_name': 'foo.something t1 Uppercase', 'table_rows': 4, 'table_size': 1234}
+            {'table_name': 'foo.t1', 'table_size': 1111},
+            {'table_name': 'foo.t2', 'table_size': 2222},
+            {'table_name': 'foo.t3', 'table_size': 3333},
+            {'table_name': 'foo.something t1 Uppercase', 'table_size': 1234}
         ]
 
         expected_output = [
-            {'table_name': 'foo.t1', 'table_rows': 1, 'table_size': 1111},
-            {'table_name': 'foo.t3', 'table_rows': 3, 'table_size': 3333},
+            {'table_name': 'foo.t1', 'table_size': 1111},
+            {'table_name': 'foo.t3', 'table_size': 3333},
         ]
 
         actual_output = utils.filter_out_selected_tables(all_schema_tables, selected_tables)