[
https://issues.jboss.org/browse/TEIID-2902?page=com.atlassian.jira.plugin...
]
Steven Hawkins commented on TEIID-2902:
---------------------------------------
So what you are asking for is a grace style join that works incrementally against sorted
sublists. I reworded the issue as it is generally not correct to think of this as pushing
the limit as pushing the limit below an inner join could result in less results than
desired.
Limit pushing over joins is done in the case of cross joins and on the left side of a left
outer join.
Also if the enhanced sort merge join is used and there is a disparity in the sizes of the
two sides, then only the smaller side will be fully fetched. The other side will be
processed incrementally. So there are several cases already when processing is more like
what you want.
More incremental join when a limit is present
---------------------------------------------
Key: TEIID-2902
URL:
https://issues.jboss.org/browse/TEIID-2902
Project: Teiid
Issue Type: Enhancement
Components: Query Engine
Affects Versions: 8.7
Reporter: Ivan Chan
Assignee: Steven Hawkins
For SQL like:
SELECT
t1.column1,
t2.column1
FROM
schema1.table1 t1,
schema2.table1 t2
WHERE
t1.column2=t2.column2
LIMIT 100
Teiid may get the whole schema1.table1 content and the whole schema2.table1 content, do
the join in memory and then, limit the resultset to 100 records.
However, I think teiid can fetch smaller set of data then do a join. If the join result
set is less than number of rows asking, keep fetching next set of data and do a join
until hits the limit number.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:
http://www.atlassian.com/software/jira