# insert overwrite table partition
df.write.mode("overwrite").partitionBy("col1","col2","col3").insertInto("db_name.table_name")