[
https://issues.apache.org/jira/browse/LUCENE-6880?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14987327#comment-14987327
]
Michael McCandless commented on LUCENE-6880:
--------------------------------------------
The javadoc for TopSuggestDocs.SuggestScoreDoc needs to be fixed (it implies
there's only one key now).
When a SuggestScoreDocs has multiple keys/contexts/scores, is there any
implication to order? Is it always sorted "best to worst" score?
I wonder if instead of 3 parallel lists, we should just have a list of
SuggestScoreDoc (as it is in trunk today) for each doc hit? In fact, this is
really like grouping? Maybe it should be a TopGroups?
It should be "fewer" not "less" in here :) : {{// This can still lead to
collecting less paths then needed...}}
> Add document oriented collector for NRTSuggester
> ------------------------------------------------
>
> Key: LUCENE-6880
> URL: https://issues.apache.org/jira/browse/LUCENE-6880
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Areek Zillur
> Assignee: Areek Zillur
> Fix For: Trunk, 5.4
>
> Attachments: LUCENE-6880.patch
>
>
> Currently NRTSuggester collects completions iteratively as they are accepted
> by the TopNSearcher, implying that a document can be collected more than
> once. In case of indexing a completion with multiple context values, the
> completion leads to {{num_context}} paths in the underlying FST for the same
> docId and gets collected {{num_context}} times, when a query matches all its
> contexts.
> Ideally, a document-oriented collector will collect top N documents instead
> of top N completions by handling the docId deduplication while collecting the
> completions. This could be used to collect n unique documents that matched a
> completion query.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
