[
https://issues.apache.org/jira/browse/HBASE-1647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12730491#action_12730491
]
ryan rawson commented on HBASE-1647:
------------------------------------
I'm afraid this patch is not good, the assumptions of the reporter are not
correct, we in fact see things in this order:
row1 / col1
row1 / col2
row2 / col1
row2 / col2
This patch removes a key piece of functionality I built into the new API that
allows us to post-facto filter a row.
The API needs to be clearer perhaps. The intent is:
- use filterRowKey() to have the earliest chance to filter an entire row based
on the _KEY_ only.
- for more complex processing, implement filterKeyValue() and set internal
state.
-- At the "End" of the KeyValues for a row, filterRow() will be called, and you
can choose, based on that state, to filter the row. Eg: if a column was
missing (which you can't know until you hit the end of the KeyValues for a
row).
If you don't use this feature, implement filterRow() { return false; }. The
JIT hotspot will take care of optimizing things.
> Filter#filterRow is called too often, filters rows it shouldn't have
> --------------------------------------------------------------------
>
> Key: HBASE-1647
> URL: https://issues.apache.org/jira/browse/HBASE-1647
> Project: Hadoop HBase
> Issue Type: Bug
> Affects Versions: 0.20.0
> Reporter: Doğacan Güney
> Fix For: 0.20.0
>
> Attachments: HBASE-1647-v2.patch, ScanBug.java, scanfilter.patch
>
>
> Filter#filterRow is called from ScanQueryMatcher#filterEntireRow which is
> called from StoreScanner.next. However, if I understood the code correctly,
> StoreScanner processes KeyValue-s in a column-oriented order (i.e. after
> row1-col1 comes row2-col1, not row1-col2). Thus, when filterEntireRow is
> called, in reality, the filter only processed (via filterKeyValue) only one
> column of a row.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
