[
https://issues.apache.org/jira/browse/PHOENIX-2724?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15360445#comment-15360445
]
Samarth Jain commented on PHOENIX-2724:
---------------------------------------
Great find,[~lhofhansl]!. One suggestion I have is to see pre-size the array
lists for rowCounts and byteCounts in GuidePostInfoBuilder.
{code}
+ private List<Long> rowCounts = new ArrayList<Long>();
+ private List<Long> byteCounts = new ArrayList<Long>();
{code}
The size of the list I believe could/should be the estimated number of guide
posts which can be computed by doing -
{code}
long numGuidePostsEstimate =
services.getProps().getLong(HConstants.HREGION_MAX_FILESIZE,
HConstants.DEFAULT_MAX_FILE_SIZE) /
services.getProps().getLong(QueryServices.STATS_GUIDEPOST_WIDTH_BYTES_ATTRIB,
QueryServicesOptions.DEFAULT_STATS_GUIDEPOST_WIDTH_BYTES);
{code}
> Query with large number of guideposts is slower compared to no stats
> --------------------------------------------------------------------
>
> Key: PHOENIX-2724
> URL: https://issues.apache.org/jira/browse/PHOENIX-2724
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.7.0
> Environment: Phoenix 4.7.0-RC4, HBase-0.98.17 on a 8 node cluster
> Reporter: Mujtaba Chohan
> Assignee: Samarth Jain
> Fix For: 4.8.0
>
> Attachments: 2724.txt, PHOENIX-2724.patch,
> PHOENIX-2724_addendum.patch, PHOENIX-2724_v2.patch
>
>
> With 1MB guidepost width for ~900GB/500M rows table. Queries with short scan
> range gets significantly slower.
> Without stats:
> {code}
> select * from T limit 10; // query execution time <100 msec
> {code}
> With stats:
> {code}
> select * from T limit 10; // query execution time >20 seconds
> Explain plan: CLIENT 876085-CHUNK 476569382 ROWS 876060986727 BYTES SERIAL
> 1-WAY FULL SCAN OVER T SERVER 10 ROW LIMIT CLIENT 10 ROW LIMIT
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
