Hi guys, am not sure if the issues below are considered to be features. I
tested it on a small set of docs (inside a single index/type), thus was
able to count what should be the outputs on particular mlt queries. The
structure of queries (taken via curl) was like this:
"query" : {
"more_like_this" : {
....
}
}
Using ES 0.20.5, the issues are:
-
min_doc_freq (probably other too) restriction is evaluated per shard;
leading to wrong omission of some docs. -
min_doc_freq (probably other too) restriction is evaluated per field;
leading to wrong inclusion of some docs. -
percent_terms_to_match restriction has strange (read nonsensical)
impacts.
Best,
M.
--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.