Hi, I have the following query that I've found in my production logs:
INFO: /select/ rows=0&start=0&f.category_id.facet.limit=-1&facet=true&facet.field=category_id&fq=product_is_active:true&fq=product_status_code:complete&fq=category_id:"1001644"&fq=attribute_id_value_en_pair:"1005585\:Wedding"&fq=merchant_id:"1000156"&qt=sti_dismax_en 0 40734 It took almost 41 seconds to run. ick! I'd like to pre-warm that query in the solrconfig.xml, but I'd like to leave off a parameter or two. For example: rows=0&start=0&f.category_id.facet.limit=-1&facet=true&facet.field=category_id&fq=product_is_active:true&fq=product_status_code:complete&fq=category_id:"1001644"&qt=sti_dismax_en I've taken off the attribute_id_value_en_pair and merchant_id fields. The query without those two parameters is never actually executed in production. Will it help the performance of the full query at all if I cache the partial query when a newSearcher is created? The reason I ask is the attribute_id_value_en_pair and merchant_id are dynamically added to the query at runtime based on results returned from our database. I can easily generate a static list of queries for each unique category_id and put them into the solrconfig.xml time in the newSearcher listener area. I'm just not sure if it will help at all if Solr uses the entire set of fields as a key into the queryResultsCache. Thanks, Brendan -- View this message in context: http://www.nabble.com/autowarm-static-queries-tf4611014.html#a13167933 Sent from the Solr - User mailing list archive at Nabble.com.
