A large start value is probably worse performing than the sort (see SOLR-1726).
Once the sort field is cached, it'll be quick from then on. Put in a warming
query in solrconfig for new and/or firstSearcher that does this sort and the
cache will be built in advance of queries at least.
Erik
On Jun 13, 2012, at 16:09 , Jack Krupansky wrote:
> Try the query without the sort to get the number of rows, then do a second
> query using a "start" equal to the number of rows. That should get you the
> last row/document.
>
> -- Jack Krupansky
>
> -----Original Message----- From: rafal.gwizd...@gmail.com
> Sent: Wednesday, June 13, 2012 3:07 PM
> To: solr-user@lucene.apache.org
> Subject: Getting maximum / minimum field value - slow query
>
> Hi, I have an index with about 9 millions of documents. Every document has an
> integer 'Id' field (it's not the SOLR document identifier) and I want to get
> the maximum value of that field.
> Therefore I'm doing a search with the following parameters
> query=*.*, sort=Id desc, rows=1
>
> <response>
> <lst name="responseHeader">
> <int name="status">0</int>
> <int name="QTime">2672</int>
> <lst name="params">
> <str name="q">*:*</str>
> <str name="rows">1</str>
> <str name="sort">Id desc</str>
> </lst>
> </lst>
> <result name="response" numFound="8747779" start="0">
> <doc>
> <str name="Uid">CRQIncident#45165891</str>
> </doc>
> </result>
> </response>
>
> The problem is that it takes quite a long time to get the response (2-10
> seconds). Why is it so slow - isn't it a simple index lookup?
>
> Best regards
> RG
>
> --
> View this message in context:
> http://lucene.472066.n3.nabble.com/Getting-maximum-minimum-field-value-slow-query-tp3989467.html
> Sent from the Solr - User mailing list archive at Nabble.com.
