[ 
https://issues.apache.org/jira/browse/DERBY-2483?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

A B updated DERBY-2483:
-----------------------

    Description: 
Pulled from the write-up for DERBY-47:

As seen in the beforeAndAfter.html document attached to DERBY-47, the 
compilation/optimization time for a query with an IN list consisting of all 
literals is far larger than it is for a similar query in which the IN list 
consists of all parameters.  Since the "probe predicate" optimization 
implemented as part of DERBY-47 should apply similarly to both types of 
queries, it would be good to figure out what it is about literals that makes 
for such a long compilation time (with respect to parameters), and to address 
that issue.

Note that this relatively excessive compilation time was an issue even before 
the changes for DERBY-47 were committed. 

  was:
Pulled from the write-up of DERBY-47:

As seen in the before-and-after.html document attached to DERBY-47, the 
compilation/optimization time for a query with an IN list consisting of all 
literals is far larger than it is for a similar in which the IN list consists 
of all parameters.  Since the "probe predicate" optimization implemented as 
part of DERBY-47 should apply similarly to both types of queries, it would be 
good to figure out what it is about literals that makes for such a long 
compilation time (with respect to parameters), and to address that issue.

Note that this relatively excessive compilation time was an issue even before 
the changes for DERBY-47 were committed. 


> Reduce compilation time for queries with large IN lists made up of literals.
> ----------------------------------------------------------------------------
>
>                 Key: DERBY-2483
>                 URL: https://issues.apache.org/jira/browse/DERBY-2483
>             Project: Derby
>          Issue Type: Improvement
>          Components: Performance
>    Affects Versions: 10.3.0.0
>            Reporter: A B
>            Priority: Minor
>
> Pulled from the write-up for DERBY-47:
> As seen in the beforeAndAfter.html document attached to DERBY-47, the 
> compilation/optimization time for a query with an IN list consisting of all 
> literals is far larger than it is for a similar query in which the IN list 
> consists of all parameters.  Since the "probe predicate" optimization 
> implemented as part of DERBY-47 should apply similarly to both types of 
> queries, it would be good to figure out what it is about literals that makes 
> for such a long compilation time (with respect to parameters), and to address 
> that issue.
> Note that this relatively excessive compilation time was an issue even before 
> the changes for DERBY-47 were committed. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to