Re: Never ending query

In article <46mbv2$fvg@xmission.xmission.com> dirk@xmission.xmission.com (Derrek Poulson) writes: >From: dirk@xmission.xmission.com (Derrek Poulson) >Subject: Re: Never ending query >Date: 25 Oct 1995 15:58:58 -0600 >jparker@hpbs3645.boi.hp.com (Jack Parker) writes: >>In talking with a cow-orker of mine it came to pass that he is having a >>similar problem with 7.1 (I don't know HOW similar yet). Apparently when >>doing a query against a view which involves a COUNT(*) and a GROUP BY, the >>optimizer goes out to lunch for 15 minutes EXACTLY of CPU time. During this >>time the entire engine is frozen to everything including sysadmin functions. >>The turbo process is sitting there chewing up clock time, put nothing else is >>going on. The sqexplain file is not written. Then 15 minutes goes by and >>the query executes - quickly too. I pointed him at the OPTCOMIND or whatever >>variable and he was just going off to try that when the problem (which has >>evidently been around for a while) went away. >>Is (was) he stirring the same pot? >Doesn't sound like it. My problem involved 5.0 and no views, counts, or >group bys. Even worse, my problem so far refuses to just go away... >>cheers >>j. >>_____________________________________________________________________________ >>Jack Parker - Hewlett Packard, DMD/IS Boise, Idaho, USA >>jparker@hpbs3645.boi.hp.com >>_____________________________________________________________________________ >> Lysdexics Untie! >>_____________________________________________________________________________ >> Any opinions expressed herein are my own and not those of my employers. >>_____________________________________________________________________________ >-- >--- >- Derrek (dirk@xmission.com) Is the query so complex that the optimiser simply has too many paths to try? If so it will keep trying until it decides it would be better off going with the best solution found so far instead of trying more paths. Ian