[
https://issues.apache.org/jira/browse/PHOENIX-2417?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15055402#comment-15055402
]
James Taylor commented on PHOENIX-2417:
---------------------------------------
Some more specifics:
- Isolate change to GuidePostsInfo as this is the in memory representation for
an entire table's worth of guideposts
- Change GuidePostsInfo {{List<byte[]> getGuidePosts()}} to instead return an
{{Iterator<byte[]>}} as it's only iterated over anyway
- Add GuidePostsInfo {{int getGuidePostsCount()}} method to replace
{{getGuidePosts().size()}} calls
- Modify GuidePostsInfo serializeGuidePostsInfo() to compress the byte[] guide
posts using prefix compress.
- Borrow or copy/paste the attached code as needed to implement the prefix
compress.
> Compress memory used by row key byte[] of guideposts
> ----------------------------------------------------
>
> Key: PHOENIX-2417
> URL: https://issues.apache.org/jira/browse/PHOENIX-2417
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: James Taylor
> Assignee: Samarth Jain
> Fix For: 4.7.0
>
>
> We've found that smaller guideposts are better in terms of minimizing any
> increase in latency for point scans. However, this increases the amount of
> memory significantly when caching the guideposts on the client. Guidepost are
> equidistant row keys in the form of raw byte[] which are likely to have a
> large percentage of their leading bytes in common (as they're stored in
> sorted order. We should use a simple compression technique to mitigate this.
> I noticed that Apache Parquet has a run length encoding - perhaps we can use
> that.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
