Robert Klemme wrote:
2006/2/16, Kittel, Martin <[EMAIL PROTECTED]>:
Hi,
isnt the engine supposed to know the size of tables?
with release 7.6 it does. For tables created in release 7.6 they are always
available. However, if you have migrated from a lower release (e.g. 7.5) these
counters have to be initialized first. With release 7.6.00.16 it's easiest to
do this by doing 'Check Data With Update' in admin mode on your database. In a
future release these counters will be initialized automatically while the
database is idle.
You can check the current values for the sizes in the system table FILES. Files
whose counters haven't been initialized yet will show NULL values for the
sizes. To get the name of the tables you can join the column FILEID with
TABLEID from TABLES (or FILEID from INDEXES for indexes).
Just wanted to make sure I didn't miss anything: this refers only to
the COUNT(*) issue, does it? Are there any other IO speed related
optimizations that make performance behavior differ significantly
between 7.5 and 7.6 (even if only for certain usage scenarios)? Thx!
Yes, :-)
we are currently working on speeding up the read i/o performance.
It will be possible to assign larger tables with few changing records
to a continous clustered storage area. This allows us to read larger
chunks of data in one i/o call, which actually improves read
performance.
... this will happen in a upcomming 7.6 release.
henrik
--
MaxDB Discussion Mailing List
For list archives: http://lists.mysql.com/maxdb
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
