Remove boxing issue on dataframes INSERT

Transpose parameter causes boxing when is set to true. This leads to memory_overhead and speed penalty.
mymarilyn · Aug 1, 2021 · f1b56bb · f1b56bb
1 parent a486eb3
commit f1b56bb
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 24 deletions.
diff --git a/clickhouse_driver/client.py b/clickhouse_driver/client.py
@@ -395,20 +395,15 @@ def query_dataframe(
         )
 
     def insert_dataframe(
-            self, query, dataframe, transpose=True, external_tables=None,
-            query_id=None, settings=None):
+            self, query, dataframe, external_tables=None, query_id=None,
+            settings=None):
         """
         *New in version 0.2.0.*
 
         Inserts pandas DataFrame with specified query.
 
         :param query: query that will be send to server.
         :param dataframe: pandas DataFrame.
-        :param transpose: whether or not transpose DataFrame before sending.
-                          This is necessary action as DataFrame can be sent in
-                          columnar form. If DataFrame is already in columnar
-                          form set this parameter to ``False``.
-                          Defaults to ``True``.
         :param external_tables: external tables to send.
                                 Defaults to ``None`` (no external tables).
         :param query_id: the query identifier. If no query id specified
@@ -423,8 +418,7 @@ def insert_dataframe(
         except ImportError:
             raise RuntimeError('Extras for NumPy must be installed')
 
-        frame = dataframe.transpose() if transpose else dataframe
-        columns = columns = [frame[col].values for col in frame]
+        columns = [dataframe[col].values for col in dataframe]
 
         return self.execute(
             query, columns, columnar=True, external_tables=external_tables,

diff --git a/tests/numpy/test_generic.py b/tests/numpy/test_generic.py
@@ -120,21 +120,6 @@ def test_insert_simple(self):
             df2 = self.client.query_dataframe('SELECT * FROM test ORDER BY a')
             self.assertTrue(df.equals(df2))
 
-    def test_insert_transposed(self):
-        n = 10
-        df = pd.DataFrame({
-            'a': range(n),
-            'b': [float(x) for x in range(n)]
-        })
-
-        with self.create_table('a Int64, b Float64'):
-            rv = self.client.insert_dataframe(
-                'INSERT INTO test VALUES', df.transpose(), transpose=False
-            )
-            self.assertEqual(rv, n)
-            df2 = self.client.query_dataframe('SELECT * FROM test ORDER BY a')
-            self.assertTrue(df.equals(df2))
-
 
 class NoNumPyTestCase(BaseTestCase):
     def setUp(self):