Hi guys, Based on my understanding, Parquet currently doesn't support pushdown filtering on the nested columns, and I might be interested in contributing this feature if it's not too difficult. I was looking into the RecordReaderImplementation class a bit, and I was wondering how important the ordering of the State[] is? Seems like we'd want to order the State[] based on columns we have filtering on to be able to skip deserializing the columns of a record we know we can toss.
Any other guidance would be appreciated as well. Thanks! *Alvin Chyan*Lead Software Engineer, Data 901 Marshall St, Suite 200, Redwood City, CA 94063 turn.com <http://www.turn.com/> | @TurnPlatform <https://twitter.com/@TurnPlatform> This message is Turn Confidential, except for information included that is already available to the public. If this message was sent to you accidentally, please delete it.
