haha, we did just that in my company, but aimed at distributed tables (like HDFS) for their partition and bucketing schemata.