initial commit

save_schema.py

@@ -0,0 +1,16 @@
+import duckdb
+
+def save_schema():
+    conn = duckdb.connect()
+    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';")
+    
+    rows = conn.execute("DESCRIBE SELECT * FROM read_parquet('s3://nearle/parquet/deliveries/*.parquet')").fetchall()
+    with open('live_schema.txt', 'w') as f:
+        for row in rows:
+            f.write(f"{row[0]} ({row[1]})\n")
+
+if __name__ == "__main__":
+    save_schema()