[ http://issues.apache.org/jira/browse/HADOOP-54?page=comments#action_12425953 ] Arun C Murthy commented on HADOOP-54: -------------------------------------
An implementation detail which I would like to bring to everyone's attention: With the 'raw' values now being a concrete object (implementing the 'ValueBytes' interface) we now have a situation in the 'base' sort of SequenceFile.Sorter where we will potentially have to store millions of 'rawValue' objects (assuming a decent sized SequenceFile with 'small' records). As it exists today the 'sort' implementation in SequenceFile.Sorter uses 'io.sort.mb' no. of bytes to 'segment' input for sorting (and later merges them). Owen suggested we also use 'no. of records' to augment the above in order to prevent situations where we might have to store millions of 'ValueBytes' objects in memory... thus we have a limit on the no. of records (i.e. ValueBytes objects) we use along with 'io.sort.mb' to segment input for sorting. Thoughts? > SequenceFile should compress blocks, not individual entries > ----------------------------------------------------------- > > Key: HADOOP-54 > URL: http://issues.apache.org/jira/browse/HADOOP-54 > Project: Hadoop > Issue Type: Improvement > Components: io > Affects Versions: 0.2.0 > Reporter: Doug Cutting > Assigned To: Arun C Murthy > Fix For: 0.6.0 > > Attachments: VIntCompressionResults.txt > > > SequenceFile will optionally compress individual values. But both > compression and performance would be much better if sequences of keys and > values are compressed together. Sync marks should only be placed between > blocks. This will require some changes to MapFile too, so that all file > positions stored there are the positions of blocks, not entries within > blocks. Probably this can be accomplished by adding a > getBlockStartPosition() method to SequenceFile.Writer. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira
