I have been trying to follow Lucene scoring across multiple searchables. And I do not see where the IDF gets normalized between searchables? (Sum DF across searchables in first half of query and use in second half of query execution to calculate right IDF across searchables.)
Lets say you have one searchable where every document has the word "foo" ( just once) and the other searchable has the the word "foo" in just 10 documents. If you run a query on "foo" your top 10 results will only come from the second searchable( all else being equal). I am guessing this is not to important since most users will be randomizing documents between searchables? I am new to the code base, and may be missing something. Wanted to check before I start implementing it. Thanks ram --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
