: By custom phrase query class I was trying to ask if it would be possible, or : even a good idea, to create a modified PhraseQuery class that is more : efficient that span queries (as I only want to use it for phrases). This : class might have multiple possible terms generated from a regex at a certain : position.
Ah! ... sorry, i missunderstood. the answer to the questions are: yes it is possible and yes it is a good idea if you don't need the full power of Spans. The answerto the question you didn't ask is: yes such a thing exists ... take a look at MultiPhraseQuery. :) QueryParser uses it when getFieldQuery returns multiple terms with a position gap of "0" ... which is perfect for things like synonys. you could have your own custom RegexPhraseQuery that can contain RegexQueries and TermQueries (instead of just Terms) and then rewrite to a MultiPhraseQuery once you rewrite the RegexQueries to a TermQueries. ...or something like that. : : : On Feb 15, 2008 5:47 PM, Chris Hostetter <[EMAIL PROTECTED]> wrote: : : > : > : I was wondering if anyone has a more efficient method for achieving : > this. : > : Would changing QueryParser.jj and developing a custom PhraseQuery class : > be a : > : good idea? Any comments would be appreciated. : > : > extending QueryParser and overriding the getPhraseQuery function to return : > your own SpanNearQueries composted of SpanRegexQueries shouldn't be any : > more efficient then modifying the QueryParser grammer. either way it's : > going to be calling some method you write that returns a SpanNearQuery. : > : > : > : > -Hoss : > : > : > --------------------------------------------------------------------- : > To unsubscribe, e-mail: [EMAIL PROTECTED] : > For additional commands, e-mail: [EMAIL PROTECTED] : > : > : -Hoss --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
