[
https://issues.apache.org/jira/browse/PHOENIX-1856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14515159#comment-14515159
]
James Taylor commented on PHOENIX-1856:
---------------------------------------
Let's not expose this in GuidePostsInfo for now as I don't think we need it.
Instead, let's minimize the change to just StatisticsCollector and
StatisticsWriter. Also, let's minimize the amount of work done in
StatisticsCollector.collectStatistics():
- Have a byte[] minKeyBytes, int minKeyOffset, and int minKeyLength and set
those when minKeyBytes is null (i.e. the first call) without instantiating an
ImmutableBytesPointer or copying any byte arrays.
- Have a byte[] maxKeyBytes, int maxKeyOffset, and int maxKeyLength and either
set that based on the last Cell or each time through the loop.
- Create a void getMinKey(ImmutableBytesWritable ptr) and a
getMaxKey(ImmutableBytesWritable ptr) that fills in the value of the ptr based
through based on the above member variables.
- Instantiate (or use already instantiated ptr) to get the min/max key from
StatisticsWriter.addStats() and do the put.add there.
> Include min and max row key for each region in stats row
> --------------------------------------------------------
>
> Key: PHOENIX-1856
> URL: https://issues.apache.org/jira/browse/PHOENIX-1856
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
> Assignee: ramkrishna.s.vasudevan
> Attachments: Phoenix-1856_1.patch
>
>
> It'd be useful to record the min and max row key for each region to make it
> easier to filter guideposts through queries.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
