Re: [Exist-open] Very slow count() query

[Roy W. <gar...@ya...>]

Jason Smith wrote:
> How about this one?  Any faster?
>
> <result>{count($hits[@speakername=$speakers[1]])}</result>
>   
0.9 secs (743 distinct speakers/6689 hits).

R.
>
> Jason Smith
> ________________________________________
> From: Roy Walter [gar...@ya...]
> Sent: Monday, July 19, 2010 2:34 PM
> To: Jason Smith
> Cc: exi...@li...
> Subject: Re: [Exist-open] Very slow count() query
>
> Jason Smith wrote:
>
> Thanks for your response.
>   
>> Have you tried rewriting the xpaths so that they read
>> $hits[@speakername=$speaker]?
>>     
> Yes, tried that. Doesn't make a big difference.
>   
>> How long does it take to just run this?
>>
>> let $hits := collection("/papers")//p[ft:query(., 'drugs')]/..
>> let $speakers := for $speaker in distinct-values($hits/@speakername)
>>     order by $speaker
>>     return $speaker
>>     
> 0.8 secs
>   
>> I suspect that the query optimizer is having problems with either the
>> original $hits, or with the syntax you've used for $speakers.
>>
>> Isolate pieces of the query, and as soon as you add something in that
>> makes it slow, that is the culprit.
>>     
> Sure, hence this:
>
>     <result>{count($hits/@speakername[.=$speakers[1]])}</result>
>
> The 'full' query does complete when run from a browser (REST). 6689
> hits; 743 distinct speakers; 68 secs. The count() part of the query is
> the bottleneck.
>
> -- Roy
>
>
>   
>> *Jason Smith
>> *
>>
>> ------------------------------------------------------------------------
>> *From:* Roy Walter [gar...@ya...]
>> *Sent:* Friday, July 16, 2010 7:16 AM
>> *To:* exi...@li...
>> *Subject:* Re: [Exist-open] Very slow count() query
>>
>> More on this. The query times out when run from oXygen. If I run the
>> query from a browser (REST) it takes more than a minute to return.
>>
>> R.
>>
>>
>> Roy Walter wrote:
>>     
>>> Consider the following in which the xquery generates a list of
>>> distinct @speakernames and then [attempts] to count the number of
>>> times each @speakername appears against a search topic:
>>>
>>>     let $hits := collection("/papers")//p[ft:query(., 'drugs')]/..
>>>     let $speakers := for $speaker in distinct-values($hits/@speakername)
>>>         order by $speaker
>>>         return $speaker
>>>     (:let $results :=
>>>     for $speaker in $speakers
>>>         return <result>{count($hits/@speakername[.=$speaker])}</result>:)
>>>     return
>>>         <result>{count($hits/@speakername[.=$speakers[1]])}</result>
>>>
>>> Getting a count of just one speaker (last line) takes approx. 5
>>> seconds. There are 700 distinct speakers (and in this case about 7000
>>> hits). What I really want to achieve is the commented part of the query.
>>>
>>> What am I doing wrong?
>>>
>>> Thanks and regards,
>>> Roy
>>> ------------------------------------------------------------------------
>>>
>>> ------------------------------------------------------------------------------
>>> This SF.net email is sponsored by Sprint
>>> What will you do first with EVO, the first 4G phone?
>>> Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
>>> ------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Exist-open mailing list
>>> Exi...@li...
>>> https://lists.sourceforge.net/lists/listinfo/exist-open
>>>
>>>