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: analysis tool vs. reality
Date Mon, 16 Aug 2010 20:20:07 GMT

: > even if you change the Lucene QUeryParser so that whitespace isn't a meta
: > character it doens't affect the underlying issue: analysis.jsp is agnostic
: > about QueryParsers.

: analysis.jsp isn't agnostic about queryparsers, its ignorant of them, and
: your default queryparser is actually a de-facto whitespace tokenizer, don't
: try to sugarcoat it.

If it makes you feel better to use the word ignorant instead of agnostic 
fine -- but i'm not suger coating anything.  analysis.jsp's "query" 
analyzer output is ignorant of all the QueryParsers that might be used at 
query time in the same way that it's "index" analyzer output is ignorant 
of the UpdateProcessors that might be used at index time -- in both 
cases it only focuses on analysis, and tells you that give input X, the 
analyzer produces output Y.

if you want to change the Lucene QueryParser then go fight that battle in 
another thread -- i'm trying to have a meaningful conversation about how 
we can better educate users about the distinction between Query Parsing 
and Analysis, and about how we can make it more clear what analysis.jsp is 
doing.

Even if you convince folks to make every change you think should be made 
to the Lucene QueryParser (again: please take that up in a seperate 
thread) it won't change the fact that people using analysis.jsp should 
understand the distinction between Query Parsing and Analysis -- unless 
you plan on getting rid of every metacharacter that the Lucene QueryParser 
uses to decide what types of Query to build (ie: '"', '-', '"', '*', '?') 
and unless you plan on forcing Solr users to only ever use that one 
QueryParser, then no matter what the Lucene QueryParser does with 
whitespace, users still need to understand the distinction between Query 
Parsing and Analysis so they don't type 'Foo*' into analysis.jsp and then 
ask why it says that will match "food" but it doesn't actually match at 
query time. (suprise suprise: Query Parsing is not the same as analysis, 
and when the QueryParser sees wildcards it doesn't use the analyzer)


-Hoss


Mime
View raw message