fix: update pyarrow references that are warning (#31)

pyarrow as of 0.17 started indicating some future warnings for
functionality that has been changed and/or relocated.  As we already
bumped the min version when fixing some of the testing here to get
around breakages related to the transitions of 0.15, it seems
reasonable to update references as well.

Author: shollyman

google/cloud/bigquery_storage_v1/reader.py

@@ -649,7 +649,7 @@ def to_dataframe(self, message, dtypes=None):
     def _parse_arrow_message(self, message):
         self._parse_arrow_schema()
 
-        return pyarrow.read_record_batch(
+        return pyarrow.ipc.read_record_batch(
             pyarrow.py_buffer(message.arrow_record_batch.serialized_record_batch),
             self._schema,
         )
@@ -658,7 +658,7 @@ def _parse_arrow_schema(self):
         if self._schema:
             return
 
-        self._schema = pyarrow.read_schema(
+        self._schema = pyarrow.ipc.read_schema(
             pyarrow.py_buffer(self._read_session.arrow_schema.serialized_schema)
         )
         self._column_names = [field.name for field in self._schema]

google/cloud/bigquery_storage_v1beta1/reader.py

@@ -607,7 +607,7 @@ def to_dataframe(self, message, dtypes=None):
     def _parse_arrow_message(self, message):
         self._parse_arrow_schema()
 
-        return pyarrow.read_record_batch(
+        return pyarrow.ipc.read_record_batch(
             pyarrow.py_buffer(message.arrow_record_batch.serialized_record_batch),
             self._schema,
         )
@@ -616,7 +616,7 @@ def _parse_arrow_schema(self):
         if self._schema:
             return
 
-        self._schema = pyarrow.read_schema(
+        self._schema = pyarrow.ipc.read_schema(
             pyarrow.py_buffer(self._read_session.arrow_schema.serialized_schema)
         )
         self._column_names = [field.name for field in self._schema]

tests/system/v1/test_reader_dataframe_v1.py

@@ -46,12 +46,12 @@ def test_read_v1(client, project_id):
 
     assert tbl.num_columns == 4
     schema = tbl.schema
-    # Use field_by_name because the order doesn't currently match that of
-    # selected_fields.
-    assert pyarrow.types.is_int64(schema.field_by_name("station_id").type)
-    assert pyarrow.types.is_float64(schema.field_by_name("latitude").type)
-    assert pyarrow.types.is_float64(schema.field_by_name("longitude").type)
-    assert pyarrow.types.is_string(schema.field_by_name("name").type)
+    # Use field with a name specifier as there may be ordering differences
+    # when selected_fields is used
+    assert pyarrow.types.is_int64(schema.field("station_id").type)
+    assert pyarrow.types.is_float64(schema.field("latitude").type)
+    assert pyarrow.types.is_float64(schema.field("longitude").type)
+    assert pyarrow.types.is_string(schema.field("name").type)
 
 
 @pytest.mark.parametrize(

tests/system/v1beta1/test_reader_dataframe_v1beta1.py

@@ -48,12 +48,12 @@ def test_read_rows_to_arrow(client, project_id):
 
     assert tbl.num_columns == 4
     schema = tbl.schema
-    # Use field_by_name because the order doesn't currently match that of
-    # selected_fields.
-    assert pyarrow.types.is_int64(schema.field_by_name("station_id").type)
-    assert pyarrow.types.is_float64(schema.field_by_name("latitude").type)
-    assert pyarrow.types.is_float64(schema.field_by_name("longitude").type)
-    assert pyarrow.types.is_string(schema.field_by_name("name").type)
+    # Use field with a name specifier as there may be ordering differences
+    # when selected_fields is used
+    assert pyarrow.types.is_int64(schema.field("station_id").type)
+    assert pyarrow.types.is_float64(schema.field("latitude").type)
+    assert pyarrow.types.is_float64(schema.field("longitude").type)
+    assert pyarrow.types.is_string(schema.field("name").type)
 
 
 @pytest.mark.parametrize(