Hi,
Is there a way to find out the rank value for each of the documents returned by the likep queries? Can I specify only the records that above certain rank point will be returned in order to improve relevancy?
Thanks!
Also note that answers are returned in rank order by default so that the best matches come out first. There's not much reason to use a cutoff value since the most relevant documents will always be first.
Thanks for the quick response! Another question is: will it improve the performance of the likep queries if I set 'likeprows' to a smaller number? any suggestions on how to improve the performance?
Lowering likeprows can improve performance, however with the default settings there wouldn't be much improvement in general by lowering it. The usual thing to check with performance are making sure you have a metamorph inverted index on the field, and that it is up to date. You can also try likepallmatch to require all terms must be present. Do you have a large number of documents that match the query?
Here is Code Section :
<sql "select id Mid, $$rank rawrank from html where StoryDescription\StoryTitle likep $sq and not(id=$id) and StoryTitle!='' and (convert('now', 'date') - Modified)/60<720 and $$rank > 300">
Catno --- $Catno Mid .... $Mid -- $rawrank <br>
</sql>
it does not work but when i do it via vortex :
<sql "select id Mid, $$rank rawrank from html where StoryDescription\StoryTitle likep $sq and not(id=$id) and StoryTitle!='' and (convert('now', 'date') - Modified)/60<720">
<if $rawrank gt 300>
Catno --- $Catno Mid .... $Mid -- $rawrank <br>
</if>
</sql>