SpanOrQuery with unreasonable high scores

[Jan Kurella <ja...@nokia.com>]

Hi,

 From what I got, the SpanOrQuery is similar to a BooleanQuery with 
Or-Clauses. It can be used in other span queries.
But compared to the BooleanQuery it has a significant difference: It is 
missing a querynormalization. Thus by the just summed up idf of all 
tokens in the query, even documents with just one hit get an extremely 
high score as mainly the idf is relevant for scoring.

To have some query normalization like in BQ one qould need to cascade 
the SpanWeight.normalize(float) down to the Spans (and its containing 
subSpans). But unfortunately neither is the Spans existing at this point 
in time nor does the Spans concept support boosts or norms.

so how do I get the SpanOrQuery best normed to be comparable to a 
BooleanQuery?

Jan


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org