[
https://issues.apache.org/jira/browse/LUCENE-1483?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael McCandless updated LUCENE-1483:
---------------------------------------
Attachment: LUCENE-1483.patch
New patch attached:
* Made IndexWriter.getSegmentCount package protected again
* Fixed StringOrdComparator to use no doubles
* Added "API is experimental" warnings to the new APIs
* Deprecated FieldValueHitQueue.getMaxScore -- I think in 3.0 we
should remove this (with Hits)?
* Tweaked StringOrdValOnDemComparator to avoid Arrays.fill in
setNextReader
* Fixed FieldValueHitQueue to actually pull the right "on demand"
comparator (it was incorrectly pulling StringOrdValComparator)
I re-ran perf tests:
||numSeg||index||sortBy||method||query||topN||hits||warm||qps||warmnew||qpsnew||pctg||
|1|simple|country|val|text|1000|2000000| 0.9| 10.0| 0.8| 11.1| 11.0%|
|1|simple|country|ord|text|1000|2000000| 0.9| 10.0| 0.7| 13.0| 30.0%|
|1|simple|country|ordval|text|1000|2000000| 0.9| 10.0| 0.7| 12.5| 25.0%|
|1|simple|country|orddem|text|1000|2000000| 0.9| 10.0| 0.7| 12.5| 25.0%|
|1|wiki|title|val|147|1000| 4984| 2.1| 740.6| 2.4| 375.7|-49.3%|
|1|wiki|title|ord|147|1000| 4984| 2.1| 740.6| 2.1| 811.7| 9.6%|
|1|wiki|title|ordval|147|1000| 4984| 2.1| 740.6| 2.1| 806.6| 8.9%|
|1|wiki|title|orddem|147|1000| 4984| 2.1| 740.6| 2.1| 207.6|-72.0%|
|1|wiki|title|val|text|1000| 97191| 2.1| 108.7| 2.4| 32.5|-70.1%|
|1|wiki|title|ord|text|1000| 97191| 2.1| 108.7| 2.1| 121.4| 11.7%|
|1|wiki|title|ordval|text|1000| 97191| 2.1| 108.7| 2.1| 119.1| 9.6%|
|1|wiki|title|orddem|text|1000| 97191| 2.1| 108.7| 2.1| 90.9|-16.4%|
|1|wiki|title|val|1|1000| 386435| 2.1| 46.2| 2.4| 12.8|-72.3%|
|1|wiki|title|ord|1|1000| 386435| 2.1| 46.2| 2.1| 58.6| 26.8%|
|1|wiki|title|ordval|1|1000| 386435| 2.1| 46.2| 2.1| 55.7| 20.6%|
|1|wiki|title|orddem|1|1000| 386435| 2.1| 46.2| 2.1| 50.1| 8.4%|
||numSeg||index||sortBy||method||query||topN||hits||warm||qps||warmnew||qpsnew||pctg||
|10|simple|country|val|text|1000|2000000| 0.8| 9.7| 0.7| 11.0| 13.4%|
|10|simple|country|ord|text|1000|2000000| 0.8| 9.7| 0.6| 13.0| 34.0%|
|10|simple|country|ordval|text|1000|2000000| 0.8| 9.7| 0.6| 12.4| 27.8%|
|10|simple|country|orddem|text|1000|2000000| 0.8| 9.7| 0.6| 12.5| 28.9%|
|10|wiki|title|val|147|1000| 4984| 12.7| 664.2| 2.5| 383.8|-42.2%|
|10|wiki|title|ord|147|1000| 4984| 12.7| 664.2| 2.1| 86.2|-87.0%|
|10|wiki|title|ordval|147|1000| 4984| 12.7| 664.2| 2.1| 104.0|-84.3%|
|10|wiki|title|orddem|147|1000| 4984| 12.7| 664.2| 2.1| 77.0|-88.4%|
|10|wiki|title|val|text|1000| 97191| 12.6| 100.4| 2.4| 33.3|-66.8%|
|10|wiki|title|ord|text|1000| 97191| 12.6| 100.4| 2.2| 80.0|-20.3%|
|10|wiki|title|ordval|text|1000| 97191| 12.6| 100.4| 2.2| 90.3|-10.1%|
|10|wiki|title|orddem|text|1000| 97191| 12.6| 100.4| 2.1| 72.1|-28.2%|
|10|wiki|title|val|1|1000| 386435| 12.7| 42.4| 2.5| 14.7|-65.3%|
|10|wiki|title|ord|1|1000| 386435| 12.7| 42.4| 2.6| 50.2| 18.4%|
|10|wiki|title|ordval|1|1000| 386435| 12.7| 42.4| 2.2| 51.3| 21.0%|
|10|wiki|title|orddem|1|1000| 386435| 12.7| 42.4| 2.2| 47.3| 11.6%|
||numSeg||index||sortBy||method||query||topN||hits||warm||qps||warmnew||qpsnew||pctg||
|100|simple|country|val|text|1000|2000000| 1.0| 8.5| 2.1| 9.2| 8.2%|
|100|simple|country|ord|text|1000|2000000| 1.0| 8.5| 0.6| 10.7| 25.9%|
|100|simple|country|ordval|text|1000|2000000| 1.0| 8.5| 0.6| 10.3| 21.2%|
|100|simple|country|orddem|text|1000|2000000| 1.0| 8.5| 0.6| 10.2| 20.0%|
|100|wiki|title|val|147|1000| 4984| 93.8| 442.8| 3.6| 238.8|-46.1%|
|100|wiki|title|ord|147|1000| 4984| 93.8| 442.8| 2.3| 19.9|-95.5%|
|100|wiki|title|ordval|147|1000| 4984| 93.8| 442.8| 2.2| 28.0|-93.7%|
|100|wiki|title|orddem|147|1000| 4984| 93.8| 442.8| 2.2| 54.1|-87.8%|
|100|wiki|title|val|text|1000| 97191| 93.4| 88.0| 3.1| 33.1|-62.4%|
|100|wiki|title|ord|text|1000| 97191| 93.4| 88.0| 2.3| 27.8|-68.4%|
|100|wiki|title|ordval|text|1000| 97191| 93.4| 88.0| 2.2| 40.9|-53.5%|
|100|wiki|title|orddem|text|1000| 97191| 93.4| 88.0| 2.2| 53.3|-39.4%|
|100|wiki|title|val|1|1000| 386435| 92.8| 41.0| 3.2| 14.7|-64.1%|
|100|wiki|title|ord|1|1000| 386435| 92.8| 41.0| 2.7| 25.3|-38.3%|
|100|wiki|title|ordval|1|1000| 386435| 92.8| 41.0| 2.2| 33.8|-17.6%|
|100|wiki|title|orddem|1|1000| 386435| 92.8| 41.0| 2.2| 42.7| 4.1%|
Haven't digested these results yet...
> Change IndexSearcher to use MultiSearcher semantics for multiple subreaders
> ---------------------------------------------------------------------------
>
> Key: LUCENE-1483
> URL: https://issues.apache.org/jira/browse/LUCENE-1483
> Project: Lucene - Java
> Issue Type: Improvement
> Affects Versions: 2.9
> Reporter: Mark Miller
> Priority: Minor
> Attachments: LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch,
> LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch,
> LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch,
> LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch, LUCENE-1483.patch,
> LUCENE-1483.patch, LUCENE-1483.patch, sortBench.py, sortCollate.py
>
>
> FieldCache and Filters are forced down to a single segment reader, allowing
> for individual segment reloading on reopen.
--
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
