[
https://issues.apache.org/jira/browse/HBASE-10262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13859523#comment-13859523
]
Feng Honghua commented on HBASE-10262:
--------------------------------------
[~anoop.hbase] Thanks for reminder. Yes, it duplicates
[HBASE-5923|https://issues.apache.org/jira/browse/HBASE-5923] by its purpose.
After going through the discussion, my understanding is it can't be achieved
gracefully since it also exposes pb type CompareType to client (according to
[~lhofhansl]'s patch and [~stack]'s suggestion).
Seems it's feasible to using CompareFilter.CompareOp as the
HTable/HTableInterface interface argument and mapping it to pb CompareType just
before making the request, as my patch does. Would [~lhofhansl] and [~stack]
check/review the patch to determine if it looks good enough to go in the trunk?
According unit-tests in TestFromClientSide are also provided. Thanks.
> provide user/client checkAndPut/checkAndDelete API with full CompareOp rather
> than only CompareOp.EQUAL
> -------------------------------------------------------------------------------------------------------
>
> Key: HBASE-10262
> URL: https://issues.apache.org/jira/browse/HBASE-10262
> Project: HBase
> Issue Type: Improvement
> Reporter: Feng Honghua
> Assignee: Feng Honghua
> Attachments: HBASE-10262-trunk_v0.patch
>
>
> currently checkAndPut/checkAndDelete provided by HTable doesn't support full
> CompareOp other than CompareOp.EQUAL, which limits user/client from doing
> more flexible checkAndPut/checkAndDelete operation. this improvement fixes
> this limit by exposing additional checkAndPut/checkAndDelete API with full
> CompareOp support.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
