[
https://issues.apache.org/jira/browse/SOLR-6319?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14085546#comment-14085546
]
Hoss Man commented on SOLR-6319:
--------------------------------
Let's put it this way...
* the comment that you quoted explicitly says "Overrequesting can help a little
here, but not as much as when sorting by count" but then does no overrequesting
at all -- which looks like a bug to me, and is at best a disconnect between the
comment and the code:
** if overrequesting can help, then why wasn't it done, or at have add an
option to do it?
** so what if it doesn't help _as much_ as with sorting by count if it still
helps? If there was _any_ overrequesting here (even if it had just been a
smaller amount then when sort=count) then the comment would match the code, and
i would have assumed it was intentional.
* i can not find any tests of distributed field faceting that combines
"sort=index + mincount>1 + limit>0" which suggests to me that the code (if not
the comment) wasn't thought through completely. (this was based on some
creative greping - if i missed an assert i'm happy to be corrected here)
* SOLR-2894 introduces new request params to allow users fine grained control
over the amount of overrequest involved if they so choose
(facet.overrequest.ratio & facet.overrequest.count) with default values that
match the current behavior of the code when facet.sort=count (limit * 1.5 + 10)
* i plan to change the existing facet.field logic such that the default
behavior for "sort=index + mincount>1 + limit>0" _will_ overrequest by default,
and respects those params (and their defaults) to decide how much over
requesting if an expert user doesn't explicitly set them.
* if you (still) don't like me calling this a bug then feel free to edit the
jira and call it WTF you want. There aren't enough hours in the day for me to
care about arguing that nit -- what i care about is making the code behavior
better by default, and having a record in jira moving forward of when the
behavior changed.
> if mincount > 1, facet.field needs to overrequest even if facet.sort=index
> --------------------------------------------------------------------------
>
> Key: SOLR-6319
> URL: https://issues.apache.org/jira/browse/SOLR-6319
> Project: Solr
> Issue Type: Bug
> Reporter: Hoss Man
> Assignee: Hoss Man
>
> Discovered this while working on SOLR-2894. the logic for distributed
> faceting ignores over requesting (beyond the user specified facet.limit) if
> the facet.sort is index order -- but the rationale for doing this falls apart
> if the user has specified a facet.mincount > 1
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
