Subject | Re: [Firebird-Architect] Re: Bi-directional indexes |
---|---|
Author | Martijn Tonies |
Post date | 2005-06-24T22:37:59Z |
Roman, others,
If FIRST n SKIP m from some largish resultset is what you're asking
for several times in a row, why isn't the largish resultset cached?
What should indices have to do with sorting? Yes, I know they're used
sometimes to do the job, but if you're getting several "pages" of rows
(not datapages) doesn't a good cache make more sense?
With regards,
Martijn Tonies
Database Workbench - tool for InterBase, Firebird, MySQL, Oracle & MS SQL
Server
Upscene Productions
http://www.upscene.com
> > Does this make any sense? Does it seem like a good idea?Doesn't it make more sense to get a useful query cache?
>
> For me that makes perfect sense.
>
> What are the gains and what are the drawbacks?
>
> Gains:
>
> - no need to define two indices for fast support of min and max
> simultaneosly;
>
> - support for scrollable cursors (?)
>
> Drawbacks:
>
> - slower index performance when many pages are empty, but this can be
> fixed by recreating index.
>
> Anything else?
If FIRST n SKIP m from some largish resultset is what you're asking
for several times in a row, why isn't the largish resultset cached?
What should indices have to do with sorting? Yes, I know they're used
sometimes to do the job, but if you're getting several "pages" of rows
(not datapages) doesn't a good cache make more sense?
With regards,
Martijn Tonies
Database Workbench - tool for InterBase, Firebird, MySQL, Oracle & MS SQL
Server
Upscene Productions
http://www.upscene.com