Schema changes on on large tables

[PR]

Hi,
I have a basic question on schema changes performed on populated tables.
Should I prepare for big roll back area and so on for both of these changes:
ALTER TABLE ADD COLUMN
ALTER TABLE ALTER COLUMN
Table contain abt. 40 mill records

Appriciate the help, as i am just DBA stand in in my project:-)

[Yukonkid]

pr*******@netscape.net (PR) wrote in message news:<d9**************************@posting.google. com>...

Hi,
I have a basic question on schema changes performed on populated tables.
Should I prepare for big roll back area and so on for both of these changes:
ALTER TABLE ADD COLUMN
ALTER TABLE ALTER COLUMN
Table contain abt. 40 mill records

Appriciate the help, as i am just DBA stand in in my project:-)

ALTER TABLE ADD COLUMN

As far this is only a change in Oracle's data repository there is
nothing to concern about rbs sizing = fast.

ALTER TABLE MODIFY COLUMN (you mean)

generally, no overhead if you keep the sizes or extend them = fast
if you shrink the sizes there's some validation that could take a
while, but nothing to concern about regarding rbs.

RBS will come into play when you have to update your new/changed
column.
I prefer several RBS split in one RBS per tablespace , enough data
files with autoextend (but with max. size limit, say 2 Gb) so the
system can choose ANY rbs to end the transaction successfully.

hope that helps

yk

[PR]

Hi,
Thanks a lot for your answer.
I checked the details of the operations (traced up the transactions),
and they are all "ALTER ADD" types. However the statements contain a
default value setting of 0,00 (Datatype DEC 2,2). I belive this
results in roll?