hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anoop Sam John (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-17191) Make use of UnsafeByteOperations#unsafeWrap(ByteBuffer buffer) in PBUtil#toCell(Cell cell)
Date Thu, 01 Dec 2016 18:17:58 GMT

    [ https://issues.apache.org/jira/browse/HBASE-17191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15712648#comment-15712648
] 

Anoop Sam John commented on HBASE-17191:
----------------------------------------

{code}
ByteBuffer row = ((ByteBufferCell) kv).getRowByteBuffer().duplicate();
2083	      row.position(((ByteBufferCell) kv).getRowPosition());
2084	      row.limit(((ByteBufferCell) kv).getRowPosition() + kv.getRowLength());
{code}
Similar code repeating for CF, Q etc.  Can we extract it into a private method? 
wrap(BB, pos, length).
You can call it with passing rowBB, cfBB etc.
kvbuilder.setRow(wrap(...));

In test can we assert CellComparator.compare (offheapKV, newOffheapKV) == 0?


> Make use of UnsafeByteOperations#unsafeWrap(ByteBuffer buffer) in PBUtil#toCell(Cell
cell)
> ------------------------------------------------------------------------------------------
>
>                 Key: HBASE-17191
>                 URL: https://issues.apache.org/jira/browse/HBASE-17191
>             Project: HBase
>          Issue Type: Improvement
>    Affects Versions: 2.0.0
>            Reporter: ramkrishna.s.vasudevan
>            Assignee: ramkrishna.s.vasudevan
>         Attachments: HBASE-17191.patch
>
>
> Since we now have support for BBs in UnsafeByteOperations#unsafeWrap(ByteBuffer) . So
for the non - java clients while creating the PB result we could avoid the copy to an temp
array. 
> Since we have a support to write to BB in ByteOutput having a result backed by Bytebuffer
should be fine. 



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

Mime
View raw message