leventov opened a new issue #4771: Another approach to capping work set size during groupBy URL: https://github.com/apache/incubator-druid/issues/4771 Rather than attempting to estimate heap footprint (see #4768), we could just make all involved data structures (lists, arrays, maps) not resizeable beyond the configured limit. The only thing that is not controllable so easily is the string data: we may group by really short or really long strings. So instead of storing them as `String` objects, we could also store them in a ByteBuffer of limited size, and operate with offsets within that ByteBuffer, where currently `String` pointers are used. It should be also more efficient, because of less indirection and less footprint (we don't need to have `String` objects and `char[]` headers in heap).
---------------------------------------------------------------- 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 --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
