empty tabl;e before inserting

delfland/ppo-insight/ppo_insight.py

@@ -203,8 +203,15 @@ def getextrapolation(extrapolateYears):
 def process_excel(filepath):
     df = pd.read_excel(filepath, engine="openpyxl")
     df.columns = df.columns.str.strip()
+
     conn, cursor = get_db_connection()
 
+    cursor.execute(
+        """
+        truncate table onderhoud
+        """
+    )
+
     cursor.execute(
         """
         CREATE TABLE IF NOT EXISTS onderhoud (
@@ -227,11 +234,19 @@ def process_excel(filepath):
     """
     )
 
+    cursor.execute(
+        "SELECT Object_code, Omschrijving_PO, Vervaldatum FROM onderhoud")
+    existing_records = set(cursor.fetchall())
+
     for _, row in df.iterrows():
         for column in row.index:
             if pd.isna(row[column]):
                 row[column] = ""
 
+        record = (row["Object code"],
+                  row["Omschrijving PO"], row["Vervaldatum"])
+
+        if record not in existing_records:
             cursor.execute(
                 """
                 INSERT INTO onderhoud 
@@ -256,6 +271,7 @@ def process_excel(filepath):
                     row["Categorie"],
                 ),
             )
+            existing_records.add(record)
 
     conn.commit()
     cursor.close()