Actual Stored Fields are no-go definitely. You can hit any kind of
forward-view index.
http://www.youtube.com/watch?v=T5RmMNDR5XI
Look at StrFieldSource, IntFieldSource. If you wonder how to access stored
fields anyway, call org.apache.lucene.index.IndexReader.document(int).
Beware of difference be
bq: Is accessing the stored data going to have a big impact on the time to
return results?
If I'm reading this right, this an anti-pattern. Consider a corpus with 10M
documents, and your query matches 1M of them. Your collector might be going
out to disk to get the stored field and decompress it 1
I'm creating a custom function (extends ValueSource). I'm generating a value
that will both be returned as a value in the hit for each doc and also be used
to sort. As I read the documentation, this is not difficult.
To determine the value for a document, I need to access the "stored" fields f
