[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15585458#comment-15585458 ] ASF subversion and git services commented on LUCENE-7493: - Commit 2b32f5d0db532723f707f0d704d7e6f6ac33 in lucene-solr's branch refs/heads/branch_6x from Mike McCandless [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=2b32f5d ] LUCENE-7493: FacetCollector.search now accepts limit=0, for getting facets but not search hits > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail.patch, LUCENE-7493-Pass.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15585456#comment-15585456 ] ASF subversion and git services commented on LUCENE-7493: - Commit 739981b6c8e6ccd60279216b320d8a25d06c70e9 in lucene-solr's branch refs/heads/master from Mike McCandless [ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=739981b ] LUCENE-7493: FacetCollector.search now accepts limit=0, for getting facets but not search hits > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail.patch, LUCENE-7493-Pass.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15585167#comment-15585167 ] Michael McCandless commented on LUCENE-7493: Excellent, the new patch looks great [~maahi333]! I'll make some minor code style fixes and commit. Thank you! > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail.patch, LUCENE-7493-Pass.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15584564#comment-15584564 ] Mahesh commented on LUCENE-7493: Thanks Michael McCandless. The above code works and test passes always now. Attached are the patch files, > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail.patch, LUCENE-7493-Pass.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[
https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15580205#comment-15580205
]
Michael McCandless commented on LUCENE-7493:
OK that failure is confusing ... you need to use this line instead to get the
facets:
{noformat}
Facets facets = getTaxonomyFacetCounts(taxo, config, facetCollector,
config.getDimConfig("b").indexFieldName);
{noformat}
Try that and see if the test always passes?
This is needed because this test's {{BeforeClass}} randomly assigns the {{b}}
field to a different index field name ...
> Support of TotalHitCountCollector for FacetCollector.search api if numdocs
> passed as zero.
> --
>
> Key: LUCENE-7493
> URL: https://issues.apache.org/jira/browse/LUCENE-7493
> Project: Lucene - Core
> Issue Type: Improvement
>Reporter: Mahesh
> Attachments: LUCENE-7493-Fail-TestCase.patch,
> LUCENE-7493-Fail-V.20.patch, LUCENE-7493-Pass-TestCase.patch,
> LUCENE-7493-Pass-V.20.patch
>
>
> Hi,
> I want to do drill down search using FacetCollection below is the code
> FacetsCollector facetCollector = new FacetsCollector();
> TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit,
> facetCollector);
> I just want facet information so I pass limit as zero but I get error
> "numHits must be > 0; please use TotalHitCountCollector if you just need the
> total hit count".
> For FacetCollector there is no way to initialize 'TotalHitCountCollector'.
> Internally it always create either 'TopFieldCollector' or
> 'TopScoreDocCollector' which does not allow limit as 0.
> So if limit should be zero then there should be a way that
> 'TotalHitCountCollector' should be initialized.
> Better way would be to provide an api which takes query and collector as
> inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15575498#comment-15575498 ] Mahesh commented on LUCENE-7493: Hey Michael McCandless thanks for pointing out. With new code there is weird behavior. Code fix and test executes as expected like before fix assertion fails and after fix test passes but the problem that I had is that sometimes test will not pass in all cases with error ' java.lang.IllegalArgumentException: dimension "b" was not indexed into field "$facets'. This is happening randomly with no fixed step to reproduce and I am not sure why :(. > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail-TestCase.patch, > LUCENE-7493-Fail-V.20.patch, LUCENE-7493-Pass-TestCase.patch, > LUCENE-7493-Pass-V.20.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[
https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15574923#comment-15574923
]
Michael McCandless commented on LUCENE-7493:
Thank you [~maahi333].
Hmm but there is a problem with your change inside {{FacetsCollector}}: I think
in the limit=0 case you will get no facet results, because you make a new
{{FacetCollector}} rather than using the {{Collector}} passed in by the user?
Can you improve the test to confirm that you do get facet results with limit=0,
which should fail, and then fix your changes in {{FacetsCollector}} and then
the test should pass?
> Support of TotalHitCountCollector for FacetCollector.search api if numdocs
> passed as zero.
> --
>
> Key: LUCENE-7493
> URL: https://issues.apache.org/jira/browse/LUCENE-7493
> Project: Lucene - Core
> Issue Type: Improvement
>Reporter: Mahesh
> Attachments: LUCENE-7493-Fail-TestCase.patch,
> LUCENE-7493-Pass-TestCase.patch
>
>
> Hi,
> I want to do drill down search using FacetCollection below is the code
> FacetsCollector facetCollector = new FacetsCollector();
> TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit,
> facetCollector);
> I just want facet information so I pass limit as zero but I get error
> "numHits must be > 0; please use TotalHitCountCollector if you just need the
> total hit count".
> For FacetCollector there is no way to initialize 'TotalHitCountCollector'.
> Internally it always create either 'TopFieldCollector' or
> 'TopScoreDocCollector' which does not allow limit as 0.
> So if limit should be zero then there should be a way that
> 'TotalHitCountCollector' should be initialized.
> Better way would be to provide an api which takes query and collector as
> inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15572184#comment-15572184 ] Mahesh commented on LUCENE-7493: Hi Michael McCandless, Added patch for failure and success case. > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493-Fail-TestCase.patch, > LUCENE-7493-Pass-TestCase.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[
https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15571964#comment-15571964
]
Michael McCandless commented on LUCENE-7493:
Thanks [~maahi333], but I was hoping for a test that *fails*, i.e. you should
change that test case to pass limit=0 to {{FacetCollector.search}} such that
when you run the test, it hits an exception.
Once we have that, then you can also go and fix {{FacetCollector.search}}
internally with your fix (to use wrapped {{TotalHitCountCollector}} when limit
is 0), at which point the test should pass.
> Support of TotalHitCountCollector for FacetCollector.search api if numdocs
> passed as zero.
> --
>
> Key: LUCENE-7493
> URL: https://issues.apache.org/jira/browse/LUCENE-7493
> Project: Lucene - Core
> Issue Type: Improvement
>Reporter: Mahesh
> Attachments: LUCENE-7493.patch
>
>
> Hi,
> I want to do drill down search using FacetCollection below is the code
> FacetsCollector facetCollector = new FacetsCollector();
> TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit,
> facetCollector);
> I just want facet information so I pass limit as zero but I get error
> "numHits must be > 0; please use TotalHitCountCollector if you just need the
> total hit count".
> For FacetCollector there is no way to initialize 'TotalHitCountCollector'.
> Internally it always create either 'TopFieldCollector' or
> 'TopScoreDocCollector' which does not allow limit as 0.
> So if limit should be zero then there should be a way that
> 'TotalHitCountCollector' should be initialized.
> Better way would be to provide an api which takes query and collector as
> inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15571680#comment-15571680 ] Mahesh commented on LUCENE-7493: Thanks Michael McCandless for helping out. Followed your steps and attached patch file. Hope its what you expect. > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > Attachments: LUCENE-7493.patch > > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[
https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15571339#comment-15571339
]
Michael McCandless commented on LUCENE-7493:
bq. Should I add test case as an attachment in this bug?
Yes please!
You can e.g. just add a new test method, e.g. {{testZeroLimit}}, to one of the
existing test cases under {{lucene/facet/src/test/...}} e.g.
{{TestDrillDownQuery}}, run it using {{ant test -Dtestcase=TestDrillDownQuery
-Dtests.method=testZeroLimit}} and it should fail at first. Then fold in your
changes above, and run the test case again, and it should pass.
Then use {{git diff HEAD}} to make a patch file that you attach back on this
issue.
> Support of TotalHitCountCollector for FacetCollector.search api if numdocs
> passed as zero.
> --
>
> Key: LUCENE-7493
> URL: https://issues.apache.org/jira/browse/LUCENE-7493
> Project: Lucene - Core
> Issue Type: Improvement
>Reporter: Mahesh
>
> Hi,
> I want to do drill down search using FacetCollection below is the code
> FacetsCollector facetCollector = new FacetsCollector();
> TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit,
> facetCollector);
> I just want facet information so I pass limit as zero but I get error
> "numHits must be > 0; please use TotalHitCountCollector if you just need the
> total hit count".
> For FacetCollector there is no way to initialize 'TotalHitCountCollector'.
> Internally it always create either 'TopFieldCollector' or
> 'TopScoreDocCollector' which does not allow limit as 0.
> So if limit should be zero then there should be a way that
> 'TotalHitCountCollector' should be initialized.
> Better way would be to provide an api which takes query and collector as
> inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15570932#comment-15570932 ] Mahesh commented on LUCENE-7493: Should I add test case as an attachment in this bug? > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[
https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15570925#comment-15570925
]
Mahesh commented on LUCENE-7493:
This is the working copy of code that I have right now.
FacetsCollector facetCollector = new FacetsCollector();
TopDocs topDocs = null;
TotalHitCountCollector totalHitCountCollector = null;
if (limit == 0) {
totalHitCountCollector = new TotalHitCountCollector();
indexSearcher.search(query, MultiCollector.wrap(totalHitCountCollector,
facetCollector));
topDocs = new TopDocs(totalHitCountCollector.getTotalHits(), new
ScoreDoc[0], Float.NaN);
} else
topDocs = FacetsCollector.search(st.searcher, filterQuery, first +
limit, facetCollector);
> Support of TotalHitCountCollector for FacetCollector.search api if numdocs
> passed as zero.
> --
>
> Key: LUCENE-7493
> URL: https://issues.apache.org/jira/browse/LUCENE-7493
> Project: Lucene - Core
> Issue Type: Improvement
>Reporter: Mahesh
>
> Hi,
> I want to do drill down search using FacetCollection below is the code
> FacetsCollector facetCollector = new FacetsCollector();
> TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit,
> facetCollector);
> I just want facet information so I pass limit as zero but I get error
> "numHits must be > 0; please use TotalHitCountCollector if you just need the
> total hit count".
> For FacetCollector there is no way to initialize 'TotalHitCountCollector'.
> Internally it always create either 'TopFieldCollector' or
> 'TopScoreDocCollector' which does not allow limit as 0.
> So if limit should be zero then there should be a way that
> 'TotalHitCountCollector' should be initialized.
> Better way would be to provide an api which takes query and collector as
> inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (LUCENE-7493) Support of TotalHitCountCollector for FacetCollector.search api if numdocs passed as zero.
[ https://issues.apache.org/jira/browse/LUCENE-7493?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15568692#comment-15568692 ] Michael McCandless commented on LUCENE-7493: Thank you [~maahi333] ... maybe you could make a test case in a patch showing the exception when you pass limit=0? I think the fix should be simple enough, basically the code you posted on the mailing list (once we debug it!)... > Support of TotalHitCountCollector for FacetCollector.search api if numdocs > passed as zero. > -- > > Key: LUCENE-7493 > URL: https://issues.apache.org/jira/browse/LUCENE-7493 > Project: Lucene - Core > Issue Type: Improvement >Reporter: Mahesh > > Hi, > I want to do drill down search using FacetCollection below is the code > FacetsCollector facetCollector = new FacetsCollector(); > TopDocs topDocs = FacetsCollector.search(st.searcher, filterQuery, limit, > facetCollector); > I just want facet information so I pass limit as zero but I get error > "numHits must be > 0; please use TotalHitCountCollector if you just need the > total hit count". > For FacetCollector there is no way to initialize 'TotalHitCountCollector'. > Internally it always create either 'TopFieldCollector' or > 'TopScoreDocCollector' which does not allow limit as 0. > So if limit should be zero then there should be a way that > 'TotalHitCountCollector' should be initialized. > Better way would be to provide an api which takes query and collector as > inputs just like 'drillSideways.search(filterQuery, totalHitCountCollector)'. -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
