|
2 | 2 | TODO list for PostgreSQL |
3 | 3 | ======================== |
4 | 4 | Current maintainer:Bruce Momjian (pgman@candle.pha.pa.us) |
5 | | -Last updated:Sun Mar13 22:00:15 EST 2005 |
| 5 | +Last updated:Mon Mar14 11:36:52 EST 2005 |
6 | 6 |
|
7 | 7 | The most recent version of this document can be viewed at the PostgreSQL web |
8 | 8 | site, http://www.PostgreSQL.org. |
@@ -680,14 +680,7 @@ Cache |
680 | 680 |
|
681 | 681 | * Consider use of open/fcntl(O_DIRECT) to minimize OS caching, |
682 | 682 | especially for WAL writes |
683 | | -* Cache last known per-tuple offsets to speed long tuple access |
684 | | - |
685 | | - While column offsets are already cached, the cache can not be used if |
686 | | - the tuple has NULLs or TOAST columns because these values change the |
687 | | - typical column offsets. Caching of such offsets could be accomplished |
688 | | - by remembering the previous offsets and use them again if the row has |
689 | | - the same pattern. |
690 | | - |
| 683 | +* -Cache last known per-tuple offsets to speed long tuple access |
691 | 684 | * Speed up COUNT(*) |
692 | 685 |
|
693 | 686 | We could use a fixed row count and a +/- count to follow MVCC |
|