Re: Performance problem

Hi Carlos,

there can be several problem respectively differences with your tables A and B:

• is table B fragmented (check DBA_EXTENTS / DBA_TABLES)?
• has table B more extents than A (check DBA_EXTENTS / DBA_TABLES)?
• is table B on "slow" filesystem as you suggested (check V$FILESTAT)?
• are the optimizer statistics valid and recent (check field LAST_ANALYZED and SAMPLE_SIZE in DBA_TABLES / DBA_INDEXES)?
• are your indexes valid (check DBA_INDEXES)?
• have the indexes on the table the same structure respectively the same fields?

...

regards
Hartmut

"Carlos Alberto" <calberto2312_at_hotmail.com> schrieb im Newsbeitrag news:72954535.0206030553.85ed69b_at_posting.google.com...
> Hi,
>
> I run a script to load a table liked this :
>
> insert into a
> select * from b
> where ...
>
> After it run :
>
> Table A has 100 million records
> Table B has 47 million records
>
> The tables have the same structure, and they are on different
> tablespaces. Tablespace which b resides is in a 'large file'
> filesystem, while the other not.
> When I count the number of records of A, it takes about 10 minutes.
> For B, it takes about 15 minutes, although it has half of records.
> Besides it, any access to table B is slower than to A, with the same
> SQL. Can you explain me why? Is there any tool (or v$) to monitor the
> IO/seconds of one datafile/tablespace? I guess it could be a physical
> problem.
>
> Thanks in advanced,
> Carlos
Received on Mon Jun 03 2002 - 10:25:14 CDT