diff --git a/scripts/helper.py b/scripts/helper.py
index 842758bb..d7e9a813 100644
--- a/scripts/helper.py
+++ b/scripts/helper.py
@@ -3,6 +3,7 @@
 
 import os
 import pandas as pd
+import pathlib
 
 
 def split_data(data_path, time_column_name, split_date):
@@ -16,28 +17,29 @@ def split_data(data_path, time_column_name, split_date):
                   if path not in (train_data_path, inference_data_path)]
 
     for file in files_list:
-        file_name = os.path.basename(file)
-        file_extension = os.path.splitext(file_name)[1].lower()
-        df = read_file(file, file_extension)
+        df = read_file(file)
 
         before_split_date = df[time_column_name] < split_date
         train_df, inference_df = df[before_split_date], df[~before_split_date]
 
-        write_file(train_df, os.path.join(train_data_path, file_name), file_extension)
-        write_file(inference_df, os.path.join(inference_data_path, file_name), file_extension)
+        file_name = os.path.basename(file)
+        write_file(train_df, os.path.join(train_data_path, file_name))
+        write_file(inference_df, os.path.join(inference_data_path, file_name))
 
     return train_data_path, inference_data_path
 
 
-def read_file(path, extension):
-    if extension == ".parquet":
+def read_file(path):
+    extension = pathlib.Path(path.lower()).suffix    
+    if extension == "parquet":
         return pd.read_parquet(path)
     else:
         return pd.read_csv(path)
 
 
-def write_file(data, path, extension):
-    if extension == ".parquet":
+def write_file(data, path):
+    extension = pathlib.Path(path.lower()).suffix    
+    if extension == "parquet":
         data.to_parquet(path)
     else:
         data.to_csv(path, index=None, header=True)