john cantu, February 06, 2017 - 4:33 pm UTC
1) I just wanted to test all three methods. I suppose that it holds true that a reorg of a table is only really needed when something abnormal has occurred like the table growing significantly from the norm. Does this also hold true with basic LOBs?
2) I just realized the confusion here. A segment shrink cascade will not be affected by any uncommitted tranx on a table. However, when a shrink is done on a LOB like so:
alter table <> modify lob(col) (shrink space COMPACT);
The session waits on "enq: TX - row lock content" unlike when performing on the table segment which isn't affected by other session's uncommitted tranx.
Thanks,
John
February 07, 2017 - 4:21 pm UTC
Thanks for getting back to us. Good info on the lob too.
dbms_redefinition
Jeff, February 07, 2017 - 10:18 pm UTC
If you really need to do table maintenance on a running system, investigate the dbms_redefinition utility package.