Refactoring of TableRecordReader (mapred / mapreduce) for reuse outside the
scope of InputSplit / RecordReader
---------------------------------------------------------------------------------------------------------------
Key: HBASE-2324
URL: https://issues.apache.org/jira/browse/HBASE-2324
Project: Hadoop HBase
Issue Type: Improvement
Reporter: Kay Kay
Attachments: HBASE-2324.patch
For the storing of tf-idf in hbase ( lucene-hbase project) we need to scan the
keys across the table and retrieve columnar values. Quite an amount of logic
can be reused from TableRecordReader for the purpose.
- Refactored TableRecordReader ( from being a protected inner class to a public
class outside )
- Created an impl class , that does the actual work, without the dependency on
hadop.mapreduce.* packages ( RecordReader) while retaining the implementation
that can be reused across libraries.
Do the same thing for .mapred. and .mapreduce. packages.
Let me know the thoughts on the same.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
