RE: Changes between 5.0.40 and 5.0.46

Were the EXPLAIN plans the same?

> -----Original Message-----
> From: Narayan Newton [mailto:nnewton@osuosl.org]
> Sent: Friday, August 24, 2007 10:43 AM
> To: internals@lists.mysql.com
> Subject: Changes between 5.0.40 and 5.0.46
>
> Hi All,
>
> Not sure if this is the correct place to ask this, and it is a fairly
> ahh...'low priority' question anyway. However, it sparked my
> interest. I
> was recently testing a pre-production setup with MySQL 5.0.46 + the
> patch for bug #30287. I was testing it against version 5.0.40 and
> noticed something odd. On the following query:
> SELECT DISTINCT m.*, r.*, u.name, u.uid FROM cvs_messages m
> INNER JOIN
> cvs_repositories r ON m.rid = r.rid INNER JOIN users u ON
> m.uid = u.uid
> ORDER BY m.created DESC LIMIT 36790, 10;
>
> The two versions had wildly different status outputs for the read_key
> and read_first handlers. Below is an example diff between the old and
> new status outputs:
> -| Handler_read_first | 77413 |
> -| Handler_read_key | 195287 |
> +| Handler_read_first | 34 |
> +| Handler_read_key | 38691 |
>
> This translates into far fewer innodb read requests:
> -| Innodb_buffer_pool_read_requests | 489063 |
> +| Innodb_buffer_pool_read_requests | 175821 |
>
> Sooo, clearly I'm not complaining...but is this a known
> improvement? The
> query plan is the same for both versions, but it seems like
> the new one
> is not doing a full index scan.
>
> --
> Narayan Newton
>
>
> --
> MySQL Internals Mailing List
> For list archives: http://lists.mysql.com/internals
> To unsubscribe:
> http://lists.mysql.com/internals?unsub=rjames@yahoo-inc.com
>
>

-- 
MySQL Internals Mailing List
For list archives: 
http://lists.mysql.com/internals
To unsubscribe:    
http://lists.mysql.com/internals?unsub=lists@pantek.com