(11/07/11 4:26), Marcus Paradies wrote:
Thanks for the helpful hints!
The debugQuery didn't work in combination with MLT for me.
If I'm using MLT in the distributed mode, how would that work? Let's assume,
I'm having 5 shards and I'm executing a MLT query which will run against all
shards. How will the response from each shard be consolidated? If I
specified mlt.count to be 10, will I get 50 or still 10 documents as a
result?
As far as I understood, scores between shards are not really comparable. So
there would be no way to retrieve the Top10 most similar documents over all
shards in a distributed search? Please correct me if I'm wrong.
Right. Unfortunately, MLT doesn't support distributed mode as I already said so.
koji
--
http://www.rondhuit.com/en/
