Doug,
It fixed the exception I was getting when trying to run a query like the ones I listed. However it if I just search for: url:store
It gets translated to: +url:url:store I haven't looked into why that might be happening but I will eventually if you don't get to it first :-) Other than that issue, I think your patch and using the QueryFilterBase class makes my url query work as I want. Thanks!
Luke
Doug Cutting wrote:
What translation do you see when you run:
bin/nutch net.nutch.searcher.Query
If you're using RawFieldQueryParser as your base, then you're probably seeing something like:
Translated: +(() () ()) +url:store/view^0.0
If so, then the problem is in part that RawFieldQueryFilter specifies a boost of 0.0, and also that RawFieldQueryFilter only works correctly when the field value is not tokenized, since it always constructs a Lucene TermQuery.
Probably we also need to fix BasicQueryFilter.java so that, when no clauses are present in the default field, nothing is added to the query. That would remove the "+(() () ())". I've added a patch which does this.
What you want to see is something more like:
Translated: +url:"store view"^2.0
To implement this you'll need some of the logic in BasicQueryFilter.java, to check if the clause is a Term or Phrase,translate it to either a Lucene TermQuery or PhraseQuery. I've attached a file called QueryFilterBase.java that should implement this correctly. It compiles, but I have not yet tested it. Tell me if it works for you.
Thanks,
Doug
------------------------------------------------------- This SF.Net email is sponsored by BEA Weblogic Workshop FREE Java Enterprise J2EE developer tools! Get your free copy of BEA WebLogic Workshop 8.1 today. http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click _______________________________________________ Nutch-developers mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/nutch-developers
