Matthias Hoys wrote:
> <Bullseye> wrote in message
> news:efo991dss2qh10tlfeanurdpjulhomsif8_at_4ax.com...
>
>>Hi!
>>
>>I have a problem - Can anyone help?
>>
>>I have a requirement to lock a table (to deny all write access to it),
>>update records in another table, and then to truncate the "locked"
>>table before allowing access to it.
>>
>>The problem is that the truncate table issues a commit (which is fine)
>>but the commit releases my lock and allows queued transactions to
>>commit BEFORE actionning the truncate (Of it seems to!).
>>
>>How can I ensure that the truncate table is executed before any other
>>transactions are committed bu other processes?
>>
>>TIA
>
>
> SELECT ... FROM ... FOR UPDATE locks the entire table.
>
>
> HTH
> Matthias
Or alternatively ... use the DBMS_LOCK package to write your own lock.
--
Daniel A. Morgan
Relational theory is not something that is simply a nice-to-have.
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)
Received on Wed May 25 2005 - 18:27:35 CDT
