refactor(dataframe): avoid chained assignment in replace operation

Author: mattbrth

influxdb_client/client/write/dataframe_serializer.py

@@ -234,7 +234,7 @@ def __init__(self, data_frame, point_settings, precision=DEFAULT_WRITE_PRECISION
 
         for k, v in dict(data_frame.dtypes).items():
             if k in data_frame_tag_columns:
-                data_frame[k].replace('', np.nan, inplace=True)
+                data_frame.replace({k: ''}, np.nan, inplace=True)
 
         self.data_frame = data_frame
         self.f = f