On Fri, Mar 18, 2016 at 12:21:13AM -0700, Alex Vandiver wrote: > On Thu, 17 Mar 2016 17:47:50 -0500 > Matt Zagrabelny <[email protected]> wrote: > > > I am curious about the benchmarks you used to make this determination? > > > In particular, the GIN fastscan option can dramatically improve search > > > performance in versions 9.4 and above. Here is a nice discussion: > > > > > > http://blog.pgaddict.com/posts/performance-since-postgresql-7-4-to-9-4-fulltext > > > > > > In addition, GIN indexes are much smaller in PostgreSQL 9.4 and above. > > > So I am curious about the data that motivated the change to GiST. > > You are quite correct; GIN indexes are superior in virtually every > way for our use case. And the documentation is unfortunately in error; > the default switched in the other direction, from GiST to GIN. See > https://github.com/bestpractical/rt/commit/e103f6da for the actual > behavior change. > Hi Alex,
I am glad to find out it is only a documentation problem. I had not had a chance to look at the actual code yet. Regards, Ken --------- RT 4.4 and RTIR Training Sessions https://bestpractical.com/training * Washington DC - May 23 & 24, 2016
