Actually it seems to me that the current behavior is unwanted.
For use cases of exact match (or lowercased), I can use string fieldtype or
keywords tokenizer.
When I want documents sponsored with only partial match I should be able to
use a whitespace tokenizer.
Current behavior prevents that as
You can sponsor more then 1 document per keyword.
And you might want to try string instead
of another FieldType. I found that textFields remove whitespace and
concatenated the tokens.
Not sure if this is intended or not.
--
View this message in context:
http://lucene.472066.n3.
