Bugs item #2549582, was opened at 2009-01-30 17:43 Message generated for change (Settings changed) made by tsheyar You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2549582&group_id=56967
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Core Group: MonetDB4 "stable" Status: Open Resolution: None >Priority: 2 Private: No Submitted By: Jan Rittinger (tsheyar) Assigned to: Nobody/Anonymous (nobody) Summary: M4: leftthetajoin estimate way to small Initial Comment: In Pathfinder generated code the PROC htordered_unique_thetajoin() calls the COMMAND leftthetajoin (located in malalgebra.mx:248). The size estimate there is 'MIN(BATcount(l), BATcount(r)' which is completely different to all other thetajoin estimates in the MonetDB code. (In my case |l|=4728, |r|=12000, and the result contains 7586008 rows. Thus using leftthetajoin costs > 60 seconds whereas using nlthetajoin---based on the same underlying primitive---takes only 2 seconds.) BTW all other thetajoins ask for an explicit estimate. As this is however a stable branch I'm not sure whether the signature should be changed. Please let me know if there is no easy and meaningful solution. Then I will just replace the leftthetajoin call in the pathfinder code base. ---------------------------------------------------------------------- >Comment By: Jan Rittinger (tsheyar) Date: 2009-01-30 21:58 Message: I replaced the leftthetajoin in the pathfinder generated code. A 'performance' fix is thus not necessary for me anymore. ---------------------------------------------------------------------- Comment By: Sjoerd Mullender (sjoerd) Date: 2009-01-30 21:36 Message: I think it would be easy enough to add an extra implementation for leftthetajoin with an additional parameter for the estimate. At this stage of the release cycle that's still an option. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2549582&group_id=56967 ------------------------------------------------------------------------------ This SF.net email is sponsored by: SourcForge Community SourceForge wants to tell your story. http://p.sf.net/sfu/sf-spreadtheword _______________________________________________ Monetdb-bugs mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
