I've read about the sort stream in v6.1 but it appears to me to break the streaming design. If it has to read all the results into memory then it's not streaming. Sounds like it could be slow and memory intensive for very large result sets. Has anyone had good results with the sort stream when there are 10M+ docs returned?
-- View this message in context: http://lucene.472066.n3.nabble.com/Specify-sorting-of-merged-streams-tp4285026p4285202.html Sent from the Solr - User mailing list archive at Nabble.com.
