Good points. At least on the limited testing we did, we were able to get the
same answer back with both executions; at least for the use cases we tested.
We are still doing more thourough testing.
This is an application that is been ported from MS SQL server to postgres and
apparently the migrat
"Efrain J. Berdecia" writes:
> In our setup it has actually worked per the explains provided making the
> query run in milliseconds instead of seconds.
To me, "work" includes "get the right answer". I do not think you
are getting the same answers that citext would normally provide.
If you don't
Thank you for the feedback.
In our setup it has actually worked per the explains provided making the query
run in milliseconds instead of seconds.
We weren't sure if this should be something that could be added natively with
future Postgres deployments.
Thanks,Efrain J. Berdecia
On Thursday
"Efrain J. Berdecia" writes:
> After attempting to use gin and gist indexes for our queries that run against
> citext columns, our team has come up with the following to make our queries
> run from 2 mins to 25ms;CREATE EXTENSION pg_trgmCREATE EXTENSION btree_gin
> --may not be needed, checking
After attempting to use gin and gist indexes for our queries that run against
citext columns, our team has come up with the following to make our queries run
from 2 mins to 25ms;CREATE EXTENSION pg_trgmCREATE EXTENSION btree_gin --may
not be needed, checking
CREATE OPERATOR CLASS gin_trgm_ops_ci
