[
https://issues.apache.org/jira/browse/DERBY-6942?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Harshvardhan Gupta updated DERBY-6942:
--------------------------------------
Attachment: DERBY-6942.diff
z13.txt
I was able to solve my problem after spending some more time on it. Attached
are two files, one is the patch which uses additional statistics for null count
values to generate accurate selectivity estimates. It build upon DERBY-6940 and
makes changes to FromBaseTable.java and TableDescriptor.java to consume the
additional statistics.
The other file is the optimizer trace after application of the patch. The extra
qualifier selectivity is changed to 0.5 representing the fraction of null
values on my dataset during last statistics update.
I also tested with multi column indexes and I was able to successfully obtain
accurate selectivity estimates for queries if this class. I encourage you to
download the patch and test the behaviour. Some more work is required to
integrate minimum and maximum column values which were also collected as part
of additional statistics.
> Utilise additional statistics for selectivity estimates.
> --------------------------------------------------------
>
> Key: DERBY-6942
> URL: https://issues.apache.org/jira/browse/DERBY-6942
> Project: Derby
> Issue Type: Sub-task
> Components: SQL
> Reporter: Harshvardhan Gupta
> Assignee: Harshvardhan Gupta
> Priority: Minor
> Attachments: DERBY-6942.diff, z12.txt, z13.txt
>
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
