[
https://issues.apache.org/jira/browse/LUCENE-1539?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12697468#action_12697468
]
Michael McCandless commented on LUCENE-1539:
--------------------------------------------
This patch still has some noise, eg the unused *Property additions to
PerfRunData, the nocommit "first" logic in ReadTask.
On DeleteTaskByPercentTask: should it delete a pctg of the undeleted
(numDocs()) docs or of the total (maxDoc()) doc space? Right now its
implementation is dangerous, eg, if I delete 5% of the index and then 10%, that
10% delete will do nothing, since the docs it deletes will fall onto the exact
docs that the 5% had deleted.
{quote}
It seems a bit awkward that DeleteByPercentTask needs to call
IR.undeleteAll before executing the deletes.
{quote}
Oh, I see. I don't think it should do that? I think it should mean "delete
XXX% of the remaining undeleted docs"?
{quote}
Also that
subsequent delete by percent calls in deletepercent.alg need to
open the latest version of the index rather than the original
(which does not have deletes)
{quote}
This seems correct? Ie the purpose of this task is "open the latest commit on
the index, delete XXX% of its undeleted docs".
{quote}
This is due to
DirectoryIndexReader.acquireWriteLock checking to insure the
latest version of the index is locked. Perhaps we can relax
this? I would rather be able to open a commit point and delete
from the reader, then flush as the latest version.
{quote}
I don't think we can relax that. This (single transaction (writer) open at
once) is a core assumption in Lucene.
> Improve Benchmark
> -----------------
>
> Key: LUCENE-1539
> URL: https://issues.apache.org/jira/browse/LUCENE-1539
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/benchmark
> Affects Versions: 2.4
> Reporter: Jason Rutherglen
> Assignee: Michael McCandless
> Priority: Minor
> Fix For: 2.9
>
> Attachments: LUCENE-1539.patch, LUCENE-1539.patch, LUCENE-1539.patch,
> LUCENE-1539.patch, sortBench2.py, sortCollate2.py
>
> Original Estimate: 336h
> Remaining Estimate: 336h
>
> Benchmark can be improved by incorporating recent suggestions posted
> on java-dev. M. McCandless' Python scripts that execute multiple
> rounds of tests can either be incorporated into the codebase or
> converted to Java.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
