[
https://issues.apache.org/jira/browse/LUCENE-4524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14301379#comment-14301379
]
Robert Muir commented on LUCENE-4524:
-------------------------------------
Thanks Alan: I benchmarked the current patch, i dont see any performance
problems:
{noformat}
Task QPS trunk StdDev QPS patch StdDev
Pct diff
IntNRQ 8.73 (5.7%) 8.21 (8.8%)
-6.0% ( -19% - 8%)
MedPhrase 261.88 (5.1%) 249.79 (4.6%)
-4.6% ( -13% - 5%)
Prefix3 218.79 (5.3%) 210.03 (6.0%)
-4.0% ( -14% - 7%)
HighPhrase 18.04 (4.2%) 17.35 (3.3%)
-3.8% ( -10% - 3%)
Wildcard 46.18 (3.3%) 44.65 (5.0%)
-3.3% ( -11% - 5%)
LowPhrase 34.95 (2.2%) 34.35 (1.5%)
-1.7% ( -5% - 1%)
HighTerm 119.48 (3.7%) 117.73 (5.2%)
-1.5% ( -10% - 7%)
MedTerm 175.53 (3.5%) 173.14 (5.0%)
-1.4% ( -9% - 7%)
LowTerm 931.30 (2.9%) 924.38 (4.5%)
-0.7% ( -7% - 6%)
HighSpanNear 143.52 (4.7%) 142.48 (3.5%)
-0.7% ( -8% - 7%)
LowSpanNear 27.97 (3.4%) 27.79 (2.6%)
-0.7% ( -6% - 5%)
AndHighLow 1167.87 (2.0%) 1161.24 (2.1%)
-0.6% ( -4% - 3%)
MedSpanNear 143.90 (4.1%) 143.30 (3.6%)
-0.4% ( -7% - 7%)
OrNotHighLow 953.80 (2.0%) 951.43 (1.7%)
-0.2% ( -3% - 3%)
LowSloppyPhrase 119.56 (3.0%) 119.79 (2.7%)
0.2% ( -5% - 6%)
Fuzzy1 107.94 (2.7%) 108.20 (3.0%)
0.2% ( -5% - 6%)
Respell 88.19 (3.3%) 88.51 (3.1%)
0.4% ( -5% - 6%)
OrNotHighMed 182.77 (2.6%) 183.48 (2.1%)
0.4% ( -4% - 5%)
MedSloppyPhrase 15.83 (4.8%) 15.91 (4.5%)
0.5% ( -8% - 10%)
Fuzzy2 66.59 (2.9%) 66.96 (3.1%)
0.5% ( -5% - 6%)
AndHighHigh 87.34 (1.9%) 88.01 (1.6%)
0.8% ( -2% - 4%)
AndHighMed 122.26 (2.0%) 123.39 (1.5%)
0.9% ( -2% - 4%)
OrNotHighHigh 48.70 (3.6%) 49.29 (4.4%)
1.2% ( -6% - 9%)
OrHighNotHigh 29.09 (3.7%) 29.44 (4.5%)
1.2% ( -6% - 9%)
OrHighLow 55.62 (7.6%) 56.35 (9.5%)
1.3% ( -14% - 19%)
OrHighNotMed 87.78 (3.9%) 88.99 (5.0%)
1.4% ( -7% - 10%)
OrHighNotLow 106.31 (4.1%) 107.84 (5.4%)
1.4% ( -7% - 11%)
OrHighMed 57.15 (7.7%) 58.06 (9.5%)
1.6% ( -14% - 20%)
OrHighHigh 26.80 (8.3%) 27.26 (10.1%)
1.7% ( -15% - 21%)
HighSloppyPhrase 13.10 (11.3%) 13.43 (12.1%)
2.5% ( -18% - 29%)
{noformat}
I will try to go thru it today and review the changes.
> Merge DocsEnum and DocsAndPositionsEnum into PostingsEnum
> ---------------------------------------------------------
>
> Key: LUCENE-4524
> URL: https://issues.apache.org/jira/browse/LUCENE-4524
> Project: Lucene - Core
> Issue Type: Improvement
> Components: core/codecs, core/index, core/search
> Affects Versions: 4.0
> Reporter: Simon Willnauer
> Fix For: 4.9, Trunk
>
> Attachments: LUCENE-4524.patch, LUCENE-4524.patch, LUCENE-4524.patch,
> LUCENE-4524.patch, LUCENE-4524.patch, LUCENE-4524.patch
>
>
> spinnoff from http://www.gossamer-threads.com/lists/lucene/java-dev/172261
> {noformat}
> hey folks,
> I have spend a hell lot of time on the positions branch to make
> positions and offsets working on all queries if needed. The one thing
> that bugged me the most is the distinction between DocsEnum and
> DocsAndPositionsEnum. Really when you look at it closer DocsEnum is a
> DocsAndFreqsEnum and if we omit Freqs we should return a DocIdSetIter.
> Same is true for
> DocsAndPostionsAndPayloadsAndOffsets*YourFancyFeatureHere*Enum. I
> don't really see the benefits from this. We should rather make the
> interface simple and call it something like PostingsEnum where you
> have to specify flags on the TermsIterator and if we can't provide the
> sufficient enum we throw an exception?
> I just want to bring up the idea here since it might simplify a lot
> for users as well for us when improving our positions / offset etc.
> support.
> thoughts? Ideas?
> simon
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
