initial commit

check_revenue.py

@@ -0,0 +1,35 @@
+import duckdb
+import pandas as pd
+
+def check_revenue():
+    conn = duckdb.connect('analytics.duckdb')
+    try:
+        conn.execute("INSTALL httpfs; LOAD httpfs;")
+        conn.execute("SET s3_region='sgp1';")
+        conn.execute("SET s3_endpoint='sgp1.digitaloceanspaces.com';")
+        conn.execute("SET s3_url_style='path';")
+        
+        # Check top riders by revenue
+        print("--- TOP 10 RIDERS BY ORDERAMOUNT ---")
+        sql = """
+        SELECT 
+            ridername, 
+            SUM(orderamount) as total_order_value,
+            COUNT(*) as total_deliveries,
+            SUM(deliveryamt) as total_delivery_fees
+        FROM read_parquet('s3://nearle/parquet/deliveries/*.parquet', union_by_name = true)
+        WHERE ridername IS NOT NULL AND ridername != ''
+        GROUP BY 1 
+        ORDER BY 2 DESC 
+        LIMIT 10
+        """
+        df = conn.execute(sql).df()
+        print(df)
+        
+    except Exception as e:
+        print(f"Error: {e}")
+    finally:
+        conn.close()
+
+if __name__ == "__main__":
+    check_revenue()