> On Dec. 11, 2014, 8:16 p.m., Brock Noland wrote: > > ql/src/test/org/apache/hadoop/hive/ql/io/TestHiveStoragePerformance.java, > > line 105 > > <https://reviews.apache.org/r/28964/diff/1/?file=789655#file789655line105> > > > > Since we don't actually use the data, should we just re-use the same > > value? That way we won't get the noise of allocating memory for new objects? > > Sergio Pena wrote: > I am just doing it the way that MR would do when getting the K,V > information. I am not sure what would happen wether we pass the same object > over and over, then the RecordReader will behave different and get different > results. I need a little extra help here to understands why next() needs the > V object.
Next takes a V object so that you can re-use the objects as opposed to allocating new ones. See "Tip 6) Reuse Writables" on this blog: http://blog.cloudera.com/blog/2009/12/7-tips-for-improving-mapreduce-performance/ - Brock ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/28964/#review64785 ----------------------------------------------------------- On Dec. 11, 2014, 7:48 p.m., Sergio Pena wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/28964/ > ----------------------------------------------------------- > > (Updated Dec. 11, 2014, 7:48 p.m.) > > > Review request for hive. > > > Bugs: HIVE-8121 > https://issues.apache.org/jira/browse/HIVE-8121 > > > Repository: hive-git > > > Description > ------- > > This is a new tool used to test ORC & PARQUET file format performance. > > > Diffs > ----- > > ql/src/test/org/apache/hadoop/hive/ql/io/TestHiveStoragePerformance.java > PRE-CREATION > > Diff: https://reviews.apache.org/r/28964/diff/ > > > Testing > ------- > > > Thanks, > > Sergio Pena > >