Hello everyone. I'm going to store csv data in a single HBase column, without breaking it up by dedicated columns for sake of space efficiency and very fast data loading to HBase. One CSV row per key. Example:
Key: identifier_abc Column1: 'aaa;bbb;ccc;ddd;eee; Key: identifier_gghh Column1: 'gh;thj;gn' Meanwhile I need to develop an HBase client that will operate with Thrift and/or REST API and the client does not need the whole bunch of CSV data for a specific key. Instead, it only needs certain fields inside CSV. I.e. field 1, 4 and 29. Thus I want to build a custom pre-processor for Thrift/REST scanner that would parse CSV inside a column and return only requested fields. Unfortunately I didn't find a way to call custom Java package that implements the pre-processing logic from Thrift or REST. Any advise would be highly appreciated. -- View this message in context: http://apache-hbase.679495.n3.nabble.com/Acessing-custom-Java-MapReduce-Filter-results-from-Thrift-REST-API-tp4087055.html Sent from the HBase Developer mailing list archive at Nabble.com.
