hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeffrey Zhong (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HBASE-10366) 0.94 filterRow() may be skipped in 0.96(or onwards) code
Date Fri, 17 Jan 2014 22:02:21 GMT

     [ https://issues.apache.org/jira/browse/HBASE-10366?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Jeffrey Zhong updated HBASE-10366:
----------------------------------

    Attachment: hbase-10366-v1.patch

Thanks All for the reviews! v1 patch adds a unit test.

[~tedyu@apache.org] The Javadoc for current hasFilterRow() is all right because new filter
implementation should follow that way in the future. I modified some comments in my previous
patch to specifically call out that 0.94 hasFilterRow() has different semantics.

> 0.94 filterRow() may be skipped in 0.96(or onwards) code
> --------------------------------------------------------
>
>                 Key: HBASE-10366
>                 URL: https://issues.apache.org/jira/browse/HBASE-10366
>             Project: HBase
>          Issue Type: Bug
>          Components: Filters
>            Reporter: Jeffrey Zhong
>            Assignee: Jeffrey Zhong
>            Priority: Critical
>             Fix For: 0.98.0, 0.96.2
>
>         Attachments: hbase-10366-v1.patch, hbase-10366.patch
>
>
> HBASE-6429 combines both filterRow & filterRow(List<KeyValue> kvs) functions
in Filter. 
> While 0.94 code or older, it may not implement hasFilterRow as HBase-6429 expected because
0.94(old) hasFilterRow only returns true when filterRow(List<KeyValue> kvs) is overridden
not the filterRow(). Therefore, the filterRow() will be skipped.
> Since we don't ask 0.94 users to update their existing filter code, the issue will cause
scan returns unexpected keyvalues and break the backward compatibility.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Mime
View raw message