Closed jenkin closed 8 years ago
thanks for the report @jenkin - I'll have a look now
I believe scores are not tracked when scrolling (See http://mail-archives.us.apache.org/mod_mbox/spark-user/201504.mbox/%3C1429349001763.b5aa674f@Nodemailer%3E), so all are reported as zero
If I directly run the same query (ie. in Sense), the result is the same, but with correct scores
but are you scrolling in the Sense GUI?
D'oh! Thank you, I missed this detail in scan&scroll feature of ES... just now I'm trying with python library and score is always zero. And no, in Sense I performed simple query, without scrolling. Ok, nevermind, I'll use simple pagination to retrieve first N results, sorry for this "not-so" issue... :)
No problem. I had forgotten about scores for scroll, I'll add a note to the docs if not already there
Hi, I'm using scan and scroll api (ie. last example here) to fetch all documents similar to others (see mlt query). All works fine, but scoring for all returned documents is always zero. If I directly run the same query (ie. in Sense), the result is the same, but with correct scores. Is there a problem during conversion from json to R data structures? Thanks! :)