[jira] [Commented] (LUCENE-8249) Add matches to exact PhraseQuery and MultiPhraseQuery

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[jira] [Commented] (LUCENE-8249) Add matches to exact PhraseQuery and MultiPhraseQuery

JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-8249?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16439293#comment-16439293 ]

Alan Woodward commented on LUCENE-8249:
---------------------------------------

Looks like a small penalty on sloppy phrases, and a slightly less small boost on exact phrases.  Or possibly just noise.
{code}
TaskQPS baseline      StdDevQPS my_modified_version      StdDev                Pct diff
        HighSloppyPhrase      589.78      (5.4%)      573.54      (7.9%)   -2.8% ( -15% -   11%)
              OrHighHigh     1174.38      (8.6%)     1146.22      (7.5%)   -2.4% ( -17% -   15%)
         MedSloppyPhrase     1328.47      (4.3%)     1302.80      (5.2%)   -1.9% ( -10% -    7%)
              AndHighLow     3138.65      (8.5%)     3087.05      (7.2%)   -1.6% ( -15% -   15%)
             LowSpanNear     1962.66      (5.3%)     1931.60      (5.8%)   -1.6% ( -12% -   10%)
                 Prefix3     1027.12      (7.8%)     1011.50      (8.0%)   -1.5% ( -16% -   15%)
                Wildcard     1842.34      (5.8%)     1821.58      (4.2%)   -1.1% ( -10% -    9%)
                PKLookup      392.44      (4.6%)      388.12      (4.6%)   -1.1% (  -9% -    8%)
   HighTermDayOfYearSort     1122.38      (6.2%)     1111.20      (7.3%)   -1.0% ( -13% -   13%)
                HighTerm     4343.88      (8.5%)     4316.70      (5.9%)   -0.6% ( -13% -   14%)
                  IntNRQ     1319.13      (2.5%)     1313.00      (2.4%)   -0.5% (  -5% -    4%)
               OrHighLow     2157.05      (4.2%)     2148.60      (4.9%)   -0.4% (  -9% -    9%)
       HighTermMonthSort     3568.59      (5.9%)     3563.38      (5.7%)   -0.1% ( -11% -   12%)
               OrHighMed     1276.34     (11.4%)     1274.61     (11.2%)   -0.1% ( -20% -   25%)
               LowPhrase     1567.69      (4.7%)     1567.03      (5.5%)   -0.0% (  -9% -   10%)
                 MedTerm     5682.98      (8.2%)     5685.03      (9.3%)    0.0% ( -16% -   19%)
             AndHighHigh     1020.12      (4.6%)     1023.48      (4.7%)    0.3% (  -8% -   10%)
         LowSloppyPhrase      885.26      (4.4%)      889.20      (5.2%)    0.4% (  -8% -   10%)
              AndHighMed     1287.27      (6.0%)     1296.46      (5.0%)    0.7% (  -9% -   12%)
                  Fuzzy1      493.78      (4.4%)      497.65      (2.9%)    0.8% (  -6% -    8%)
                  Fuzzy2       83.87     (20.0%)       85.02     (18.4%)    1.4% ( -30% -   49%)
                 Respell      391.63      (4.6%)      397.30      (4.1%)    1.4% (  -6% -   10%)
                 LowTerm     6098.16      (6.0%)     6202.87      (5.4%)    1.7% (  -9% -   13%)
            HighSpanNear      773.18     (10.9%)      786.87      (8.4%)    1.8% ( -15% -   23%)
             MedSpanNear      937.52      (6.1%)      960.49      (4.2%)    2.4% (  -7% -   13%)
              HighPhrase     1035.86      (3.8%)     1101.79      (4.9%)    6.4% (  -2% -   15%)
               MedPhrase      997.89      (7.2%)     1068.68      (5.0%)    7.1% (  -4% -   20%)
{code}

> Add matches to exact PhraseQuery and MultiPhraseQuery
> -----------------------------------------------------
>
>                 Key: LUCENE-8249
>                 URL: https://issues.apache.org/jira/browse/LUCENE-8249
>             Project: Lucene - Core
>          Issue Type: Improvement
>            Reporter: Alan Woodward
>            Assignee: Alan Woodward
>            Priority: Major
>         Attachments: LUCENE-8249.patch, LUCENE-8249.patch
>
>
> ExactPhraseScorer can be rejigged fairly easily to expose a MatchesIterator



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]