Add clean_mshr19_NV.py

clean_mshr19_NV.py

@@ -0,0 +1,25 @@
+import pandas as pd
+import glob
+import os
+
+# Collect all .xls and .xlsx files in the current directory
+excel_files = glob.glob("*.xls") + glob.glob("*.xlsx")
+
+for file_path in excel_files:
+    print(f"Processing: {file_path}")
+    # Read the first sheet
+    df = pd.read_excel(file_path, sheet_name=0, header=None)
+    
+    # Remove the first 9 rows
+    df = df.iloc[9:, :]
+    
+    # Remove rows where column A has "TỔNG"
+    df = df[df.iloc[:, 0] != "TỔNG"]
+    
+    # Construct output file name
+    base_name = os.path.splitext(file_path)[0]  # e.g. "data"
+    output_file = f"{base_name}_cleaned.xlsx"
+    
+    # Save the result
+    df.to_excel(output_file, index=False, header=False)
+    print(f" --> Cleaned file saved as: {output_file}")