[jira] [Comment Edited] (HBASE-14841) Allow Dictionary to work with BytebufferedCells

["ramkrishna.s.vasudevan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-14841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15081000#comment-15081000 ] 

ramkrishna.s.vasudevan edited comment on HBASE-14841 at 1/4/16 11:07 AM:
-------------------------------------------------------------------------

Yes that is what I was saying in my comment. Though there is one benefit now. When the Tag is not in dictionary when we write the tag as a whole to the OutputStream we avoid the copy now with this patch (previously it was copying to a byte array).
{code}
    if (dictIdx == Dictionary.NOT_IN_DICTIONARY) {
      out.write(Dictionary.NOT_IN_DICTIONARY);
      StreamUtils.writeRawVInt32(out, length);
      ByteBufferUtils.writeByteBuffer(out, data, offset, length);
    } else {
      StreamUtils.writeShort(out, dictIdx);
{code}


was (Author: ram_krish):
Yes that is what I was saying in my comment. Though there is one benefit now. When the Tag is not in dictionary when we write the tag as a whole to the OutputStream we avoid the copy now with this patch (previously it was copying to a byte array).
{code}
{code}

> Allow Dictionary to work with BytebufferedCells
> -----------------------------------------------
>
>                 Key: HBASE-14841
>                 URL: https://issues.apache.org/jira/browse/HBASE-14841
>             Project: HBase
>          Issue Type: Sub-task
>          Components: regionserver, Scanners
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>         Attachments: HBASE-14841.patch
>
>
> This is part of HBASE-14832 where we need to ensure that while BBCells are getting compacted the TagCompression part should be working with BBCells.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)