Steve, Have a look at stream-lib and its QDigest implementation for mergable percentiles. That's what we currently use in SPM....until Ted Dunning shares his lean implementation ...
Otis Solr & ElasticSearch Support http://sematext.com/ On Nov 19, 2013 7:23 AM, "Steven Bower" <smb-apa...@alcyon.net> wrote: > Would love to see it get onto trunk and people start to play with it > more.. Between now and the new year I have some "down time" so if there are > major changes/fixes/enhancements I'll have more time to make them.. > > Also note that this stuff doesn't work multi-shard and I've not had time > to test it... Some parts will be very hard to work multi-shard (eg > median/percentile calculation).. Will try and get this tested in the near > future. > > steve > > > On Tue, Nov 19, 2013 at 5:42 AM, Shalin Shekhar Mangar < > shalinman...@gmail.com> wrote: > >> +1 for committing it. Progress over perfection. >> >> On Thu, Nov 7, 2013 at 7:20 PM, Erick Erickson <erickerick...@gmail.com> >> wrote: >> > All: >> > >> > I'm a little overwhelmed by this thing. It's 500K and currently fails >> > precommit with javadoc warnings. I can fix those bits with some >> > boilerplate... >> > >> > But, how to proceed? I've no hope of competently reviewing 500K of >> changes >> > (some if it is documentation, true). But it obviously have lots and >> lots of >> > work put in to it, and I'd like to get moving on getting this checked >> in. >> > >> > I'm looking for some consensus here on whether just fixing up the >> javadocs >> > and running the test suite is enough to at least check it in to trunk. >> From >> > there, letting it bake for a week or two and merging into 4.x (after >> 4.6 is >> > cut) would be the next step. >> > >> > It's new functionality, so it doesn't seem _too_ dangerous. >> > >> > See, if people don't raise objections, I can spread the blame if things >> go >> > wrong :) >> >> >> >> -- >> Regards, >> Shalin Shekhar Mangar. >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org >> For additional commands, e-mail: dev-h...@lucene.apache.org >> >> >