Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Jim C. Nasby
On Thu, Oct 12, 2006 at 01:58:22PM -0700, Josh Berkus wrote: > > Unless you've got a time machine or a team of coders in your back > > pocket, I don't see how the planner will suddenly become perfect in > > 8.4... > > Since you're not a core code contributor, I really don't see why you > continue

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Tom Lane
"Jim C. Nasby" <[EMAIL PROTECTED]> writes: > I completely agree that it's much better *in the long run* to improve > the planner and the statistics system so that we don't need hints. But > there's been no plan put forward for how to do that, which means we also > have no idea when some of these pr

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Bucky Jordan
> I completely agree that it's much better *in the long run* to improve > the planner and the statistics system so that we don't need hints. But > there's been no plan put forward for how to do that, which means we also > have no idea when some of these problems will be resolved. If someone > comes

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Josh Berkus
Jim, > Well, that's not what I said (my point being that until the planner and > stats are perfect you need a way to over-ride them)... but I've also > never said hints would be faster or easier than stats modification (I > said I hope they would). Yes, you did. Repeatedly. On this and other th

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Josh Berkus
Andreas, > I think we need to more precisely define the problems of our system with > point in time statistics > > -- no reaction to degree of other concurrent activity > -- no way to react to abnormal skew that only persists for a very short > duration > -- too late reaction to changing distribut

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Jim C. Nasby
On Fri, Oct 13, 2006 at 03:57:23PM -0700, Josh Berkus wrote: > Jim, > > > Well, that's not what I said (my point being that until the planner and > > stats are perfect you need a way to over-ride them)... but I've also > > never said hints would be faster or easier than stats modification (I > > s

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Alvaro Herrera
Jim C. Nasby wrote: > So, does anyone out there have a plan for how we could give user's the > ability to control the planner at a per-table level in 8.3 or even 8.4? Per-table level? Some of the problems that have been put forward have to do with table combinations (for example selectivity of j

Re: [HACKERS] [PERFORM] Hints proposal

2006-10-13 Thread Tom Lane
"Jim C. Nasby" <[EMAIL PROTECTED]> writes: > Let me clarify, because that's not what I meant. Right now, there's not > even a shadow of a design for anything else, and this is a tough nut to > crack. I think you are not exactly measuring on a level playing field. On the textually-embedded-hints s