[JBoss JIRA] (TEIID-4812) Processor is not exiting the join as soon as it's determined one side contains 0 rows
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-4812?page=com.atlassian.jira.plugin... ]
Steven Hawkins resolved TEIID-4812.
-----------------------------------
Fix Version/s: 9.3
9.2.2
Resolution: Done
Added several changes related to large temp table reads and 0 rows from one side of the join. We'll now look to honor the timeslice in sorting (since we can pull without blocking from temp tables) and not push non-key based sorting to temp tables (as that would be processed without blocking).
> Processor is not exiting the join as soon as it's determined one side contains 0 rows
> -------------------------------------------------------------------------------------
>
> Key: TEIID-4812
> URL: https://issues.jboss.org/browse/TEIID-4812
> Project: Teiid
> Issue Type: Bug
> Components: Server
> Affects Versions: 8.12.8.6_3, 8.7.11.6_2
> Reporter: Debbie Steigner
> Assignee: Steven Hawkins
> Fix For: 9.3, 9.2.2
>
>
> There is an issue with the processor during a join, it doesn't exit the join as soon as it's determined that one side of the join returns 0 rows.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 9 months
[JBoss JIRA] (TEIID-4810) Translator Caching
by Steven Hawkins (JIRA)
[ https://issues.jboss.org/browse/TEIID-4810?page=com.atlassian.jira.plugin... ]
Steven Hawkins reassigned TEIID-4810:
-------------------------------------
Assignee: (was: Steven Hawkins)
See TEIID-1598 for the server side logic, which is also covered in the docs - https://teiid.gitbooks.io/documents/content/dev/Caching_API.html
In general there's just one method that the translator must implement to interact with the result caching.
This issue is covering elevating that something that is built-in / configurable, and exposed in the UI. I would recommend pursuing this as a low priority as the higher level caching options generally offer more reuse and greater time savings.
> Translator Caching
> ------------------
>
> Key: TEIID-4810
> URL: https://issues.jboss.org/browse/TEIID-4810
> Project: Teiid
> Issue Type: Enhancement
> Components: Misc. Connectors, Query Engine
> Affects Versions: 9.3
> Reporter: Van Halbert
>
> we also have a whole different level of caching for custom translators
> we could expand on that
> we could cache on any source query
> that’s not something we’ve talked about putting any higher level UI around
> this would be a development and documentation effort
> this is different than the whole model materialisation
> the difference is that this is at a source query level
> vs when you are using cache hints and extension metadata
> we don’t have anything for the design-time user
> that’s another feature set around caching that we could put some exposure on it
> otherwise it’s not really exposed
> the idea there is that if you want to do some higher level caching
> you can build that into your models or your user query
> we could turn it on and off without having to use a custom translator
> MMX had a feature like that but it was convoluted because it was mixed up with the user level query caching and they had to be enabled together
> Attachments
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 9 months