kevinjmh commented on issue #3444: [CARBONDATA-3581] Support page level bloom filter URL: https://github.com/apache/carbondata/pull/3444#issuecomment-556949448 > > Store size increased 1% when I set a string column and a bigint column to build page bloom. > > Query performance is about 33.3% better in condition of no sort for point query, since many pages are skipped. > > What was the size of the data ? and did you benchmark with local sort also ? About 88milion rows. I did compare with local sort. Local sort only ask for several tasks but no sort ask a lot, so local sort in quite fast. In a case, local sort cost 0.9s, no sort cost 3.0s and no sort with page bloom cost 2.0s.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
