lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Hostetter <hossman_luc...@fucit.org>
Subject Re: Short DismaxRequestHandler Question
Date Wed, 12 May 2010 20:11:21 GMT

: I got a fundamental understanding question that Mike's posting did not
: answer:
: You say "q=apple iPhone & qf=title^5 manufacturer & mm=100%" is correct.
: That means:
: title: "iphone" -> matches "iphone" but not "apple"
: manufacturer: "apple" -> matches "apple" but not "iphone"
: According to the query, at least 100% of the query-terms matched the doc.
: So far, so good.
: 
: If I would define "apple" as a stopword for title, what would be the
: problem? What is the difference for the handler? In this example, there
: would never be a hit on "apple", regardless whether I say it's a stopword or
: not.

there is no problem, there may in fact be a "hit" on the word "apple" 
in the "manufacturer" field -- that is compleltey irelevant to wether 
"apple" is a stopword in the title field.  what is relevant is:
 * the word "iphone" is valid in both the title and manufacturer fields
   - therefore we have a clause that will match "iphone" in either field
 * the word "apple" is valid in the manufacturer field
   - therefore we have a clause that will match "apple" in the 
     manufacturer field
 * mm=100% and we have 2 clauses
   - therefore a minimum of 2 clauses must match

... therefore we must have a match on "apple" in the manufacture field in 
addition to matching "iphone" on one of hte other fields.

: I am not sure, whether this answers my question.
: Let me try to explain what I've understood: title says "apple" is invalid,
: which means that "apple" is assumed to be invalid for the whole query, and
: so the query can't match 100%, although it's valid for "manufacturer". Am I
: correct?

No ... not at all -- as i've said before: you have to *try* it and look at 
the debug information.  if it's not clear from a textual description of 
one example, then it won't be any clearer from textual descriptions of 100 
examples -- no amount of verbal explanation are going to be as useful as 
looking at the query structure produced for multiple examples (with and 
w/o stopwords, with and with/out ngrams, etc...)


-Hoss


Mime
View raw message