![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
[snip] Whilst waiting for Ingres to fix the Issue, I remembered a conversation I had with an Ingres engineer at, IIRC the UKIUA at Blackpoll. His idea was that in extreme conditions where Ingres can’t help but produce a ‘bad’ plan, a method could be employed to make Ingres evaluate the best flattened plan AND the best un-flattened plan and choose the best overall plan i.e. compile two plans and pick the best one. |
#3
| |||
| |||
|
|
On Jan 15, 2010, at 8:06 AM, rthdavid wrote: That might have been me ... Karl |
#4
| |||
| |||
|
|
On 15 Jan, 13:43, Karl Schendel <schen... (AT) kbcomputer (DOT) com> wrote: On Jan 15, 2010, at 8:06 AM, rthdavid wrote: That might have been me ... Karl Yes. It was you. :-) The Blackpool UKIUA was one of the best jolly's ever. Only CA World 1999 beats it... |
![]() |
| Thread Tools | |
| Display Modes | |
| |