[
https://issues.apache.org/jira/browse/PHOENIX-1287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14374815#comment-14374815
]
James Taylor commented on PHOENIX-1287:
---------------------------------------
In Phoenix, an byte[] with length of zero is treated as null. This is the same
way Oracle functions.
If all the tests pass, this is probably a good chunk of work to check-in, yes?
I'd be interested in seeing how performance is impacted by using the joni
library version regular Java regex. To measure that, you can use our
bin/performance.py script to generate ~10M rows, then measure the time for a
LIKE with and without the config property that controls whether or not the new
implementation is used. I'd try with a query that does a FULL TABLE SCAN over
the data.
> Use the joni byte[] regex engine in place of j.u.regex
> ------------------------------------------------------
>
> Key: PHOENIX-1287
> URL: https://issues.apache.org/jira/browse/PHOENIX-1287
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
> Assignee: Shuxiong Ye
> Labels: gsoc2015
>
> See HBASE-11907. We'd get a 2x perf benefit plus it's driven off of byte[]
> instead of strings.Thanks for the pointer, [~apurtell].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
