Re: doubt on table scan

<yong321_at_yahoo.com> a écrit dans le message de news: 1129736026.617967.93940_at_o13g2000cwo.googlegroups.com...
| Hi, Michel,
|
| The query result looks like coming from a parallel query coordinator.
| Was the table altered parallel <a number probably larger than 5> by any
| chance? Can you provide the full SQLs including the SQL that creates
| the tablespace? But I think the most important is how the table is
| created (or altered). What's the block size of the database or
| tablespace, and database version? Thanks.
|
| Yong Huang
|

Yong Huang, AnySQL,

You're right, i cheated using a parallel 3 (just 3) on my table but noparallel was not in the OP question. This was just to show that they are many and many cases and a general question like his is meaningless. I can build several other tests with constraints, indexes or other things that make the result differs from just "extent allocation order".

For instance, just adding a PK on ID and then you get the rows in order of ID.

Regards
Michel Cadot Received on Wed Oct 19 2005 - 11:32:41 CDT