Jonathan and Martin,
Thanks for replying. You both pointed out the same possibility which I
agree.
Allen
"Martin Drautzburg" <martin.drautzburg_at_web.de> wrote in message
news:87znmz2tdq.fsf_at_web.de...
> "Yun Guan" <yguan_at_houston.rr.com> writes:
>
> > I have one table anaylyzed using "compute statistics". It has 2774
Blocks,
> > which spans over 3 extents, chain_cnt=0, block_size 8192,
ave_row_len=47. I
> > decided to move all data into one larger extent. So I dropped the table,
> > recreate the table with initial extent = 22M (> 2774 x 8/1024) with all
> > other storage parameters unchanged. After re-analyzing it, the table has
> > 3015 blocks and spans over 2 extents. I could not figure out why the
block
> > number increases with chain_cnt =0. It spans over 2 extents, not 1
extent
> > as I expected.
>
> Could it be that this is a PCTFREE thing. The original Table might
> have grown due to updates without new blocks being allocated using up
> some of the PCTFREE margin. When the same data is inserted into
> another table oracle tries to leave a full amount of PCTFREE free.
Received on Wed Apr 09 2003 - 22:51:32 CDT
