[
https://issues.apache.org/jira/browse/PHOENIX-1453?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14262350#comment-14262350
]
James Taylor commented on PHOENIX-1453:
---------------------------------------
Thanks for the continued work on this, [~ramkrishna]. It's getting close. This
patch shouldn't get committed to the 4.2 branch, though. Once you make the
below changes, please make sure the above exception is not happening with the
4.0 branch.
Let's not put byteCount and rowCount in parallel arrays. Instead, let's keep a
single byteCount and rowCount *per column family* in the GuidePostInfo that
goes back to the client. To combine GuidePostRegionInfo with the GuidePostInfo,
you'll just sum the guidePostRegionInfo.rowCount with the
guidePostInfo.rowCount (and same with byteCount). That'll be the most useful
and simplest for now. This will require a few changes to the protobufs, so
let's do your suggestion of having a top level PGuidePost proto. So this will
have for each column family List<byte[]> guideposts, long byteCount, and long
rowCount.
Once you do the above, this may get fixed, but this doesn't look correct.
{code}
+ // CF A alone has double the bytecount because one is due to the
default qualifier _0 and
+ // other due to the CF A
+ long[] byteCountArr = new long[]{12120, 5540, 6652, 6652};
{code}
Also, please make sure that an old client doesn't break when it reads the stats
from a new server.
> Collect row counts per region in stats table
> --------------------------------------------
>
> Key: PHOENIX-1453
> URL: https://issues.apache.org/jira/browse/PHOENIX-1453
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: James Taylor
> Assignee: ramkrishna.s.vasudevan
> Attachments: Phoenix-1453.patch, Phoenix-1453_1.patch,
> Phoenix-1453_10.patch, Phoenix-1453_13.patch, Phoenix-1453_2.patch,
> Phoenix-1453_3.patch, Phoenix-1453_7.patch, Phoenix-1453_8.patch
>
>
> We currently collect guideposts per equal chunk, but we should also capture
> row counts. Should we have a parallel array with the guideposts that count
> rows per guidepost, or is it enough to have a per region count?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
