Re: Query performance issue while using EdgeNGram
Hmmm. "find evicted docs"? If you mean find out how many docs are deleted, look on the admin "schema browser" page and the difference between MaxDoc and NumDocs is the number of deleted documents. You say "for some queries the QTime is more than 8 secs". What happens if you re-run that query a bit later? The reason I ask is if you're not warming the cache that that particular query uses, you may be seeing cache loading time here. Look at the admin stats page, especially for evictions. It's also possible that your caches are being reclaimed for some queries and you're seeing response time spikes when the caches are re-loaded. Best Erick On Wed, Dec 22, 2010 at 7:10 AM, Shanmugavel SRD wrote: > > 1) Thanks for this update. I have to use 'WhiteSpaceTokenizer' > 2) I have to suggest the whole query itself (Say name or title) > 3) Could you please let me know if there is a way to find the evicted docs? > 4) Yes, we are seeing improvement in the response time if we optimize. But > still for some queries QTime is more than 8 secs. It is a 'Blocker' for us. > Could you please suggest any to reduce the QTime to 1 secs. > -- > View this message in context: > http://lucene.472066.n3.nabble.com/Query-performance-issue-while-using-EdgeNGram-tp2097056p2130751.html > Sent from the Solr - User mailing list archive at Nabble.com. >
Re: Query performance issue while using EdgeNGram
1) Thanks for this update. I have to use 'WhiteSpaceTokenizer' 2) I have to suggest the whole query itself (Say name or title) 3) Could you please let me know if there is a way to find the evicted docs? 4) Yes, we are seeing improvement in the response time if we optimize. But still for some queries QTime is more than 8 secs. It is a 'Blocker' for us. Could you please suggest any to reduce the QTime to 1 secs. -- View this message in context: http://lucene.472066.n3.nabble.com/Query-performance-issue-while-using-EdgeNGram-tp2097056p2130751.html Sent from the Solr - User mailing list archive at Nabble.com.
Re: Query performance issue while using EdgeNGram
ypeahead name="start">0100
> 4typeahead name="start">0100
> 5typeahead name="start">0100
> 6typeahead name="start">0100
> 7typeahead name="start">0100
> 8typeahead name="start">0100
> 9typeahead name="start">0100
>
>
>
>
> atypeahead name="start">0100
> btypeahead name="start">0100
> ctypeahead name="start">0100
> dtypeahead name="start">0100
> etypeahead name="start">0100
> ftypeahead name="start">0100
> gtypeahead name="start">0100
> htypeahead name="start">0100
> itypeahead name="start">0100
> jtypeahead name="start">0100
> ktypeahead name="start">0100
> ltypeahead name="start">0100
> mtypeahead name="start">0100
> ntypeahead name="start">0100
> otypeahead name="start">0100
> ptypeahead name="start">0100
> qtypeahead name="start">0100
> rtypeahead name="start">0100
> stypeahead name="start">0100
> ttypeahead name="start">0100
> utypeahead name="start">0100
> vtypeahead name="start">0100
> wtypeahead name="start">0100
> xtypeahead name="start">0100
> ytypeahead name="start">0100
> ztypeahead name="start">0100
> 0typeahead name="start">0100
> 1typeahead name="start">0100
> 2typeahead name="start">0100
> 3typeahead name="start">0100
> 4typeahead name="start">0100
> 5typeahead name="start">0100
> 6typeahead name="start">0100
> 7typeahead name="start">0100
> 8typeahead name="start">0100
> 9typeahead name="start">0100
>
>
>false
>20
>
>
>
>
> json
> name,score
> score desc
> true
>
>
>
> multipartUploadLimitInKB="2048" />
>
>
>
> default="true">
>
> explicit
> 10
>
>
>
>
> class="org.apache.solr.handler.component.SearchHandler">
>
> explicit
>
>
> class="solr.SpellingQueryConverter"/>
> class="org.apache.solr.handler.component.QueryElevationComponent" >
>string
>elevate.xml
>
> class="org.apache.solr.handler.component.SearchHandler" startup="lazy">
>
> explicit
>
>
> elevator
>
>
> >
> >
> startup="lazy" />
> class="org.apache.solr.handler.admin.AdminHandlers" />
>
>
> standard
> solrpingquery
> all
>
>
>
>default="true">
>
> 100
>
>
>class="org.apache.solr.highlight.RegexFragmenter">
>
> 70
> 0.5
> [-\w ,/\n\"']{20,200}
>
>
>default="true">
>
>
>
>
>
>
> class="org.apache.solr.request.XMLResponseWriter" default="true"/>
> class="org.apache.solr.request.JSONResponseWriter"/>
> class="org.apache.solr.request.PythonResponseWriter"/>
> class="org.apache.solr.request.RubyResponseWriter"/>
> class="org.apache.solr.request.PHPResponseWriter"/>
> class="org.apache.solr.request.PHPSerializedResponseWriter"/>
> class="org.apache.solr.request.XSLTResponseWriter">
>5
>
>
>
> schema.xml
>
>
>
>
>
>
>
> pattern="([^a-z0-9])"
> replacement="" replace="all" />
> maxGramSize="100"
> minGramSize="1" />
>
>
>
>
>
> pattern="([^a-z0-9])"
> replacement="" replace="all" />
> pattern="^(.{20})(.*)?" replacement="$1" replace="all" />
>
>
>
>
>
>
>
>stored="false"/>
>
> id
> autosuggest
>
>
> --
> View this message in context:
> http://lucene.472066.n3.nabble.com/Query-performance-issue-while-using-EdgeNGram-tp2097056p2097056.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
