Closed atarkowska closed 7 years ago
@pwalczysko you can test it not on http://10.0.51.150/web/webclient/?experimenter=-1
Really like the behaviour a lot, much much better than previously, now I really know which "cdc"s are there.
The only question is ordering of the first results matchin value% - you say they are sorted by lenght.
How does this work in case of the below example - obviously the search for ccdc
have brought plenty of resultes matching value% with the same length, but of course I would ask myself why is for example CCDC3 listed after CCDC7 ?
I think this is really a much smaller issue than the one which this PR solved, so happy to merge.
it is sorted now by both functions. I don't see much difference in the performance in psql.
@joshmoore if that land in demo3, any cached under /var/cache/nginx/omeromapr
containing autocomplete
has to be invalidated
Also the second sorting works as expected.
As @aleksandra-tarkowska remarks, I do not see any worsening of performance worth talking about. Ready to merge fmpov.
Thanks, all. @aleksandra-tarkowska : how do you want to handle the version bump?
sorry, forgot about it, done
B-eautiful.
This fixes autocomplete search that will results in returning first results matching value% (starting from the given phrase) then everything that match %value% (contains given phrase)
To test check:
cdc
,ab
,cell
acid
first results matching value% will be sorted by lenght, rest of the results matching %value% will be sorted alphabetical
cc: @pwalczysko