| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
Home -> Community -> Usenet -> c.d.o.misc -> Use of COMMIT WORK;
According to the SQL-92 standard, the command COMMIT WORK will
commit all outstanding transactions, and subsequent ROLLBACK
commands can only roll back to the most recent COMMIT.
So far so good. But what happens in practice when you're
entering SQL "conversationally" through the Oracle text window
(Sorry I forget what Oracle call it), or similarly for Tandem SQL
with sqlci?
How does the system know you want to use atomic transactions BEFORE you do the first COMMIT WORK? Do you have to do a COMMIT at the start to set up a voidable transaction, or does it assume you always do, and not commit anything at all until you tell it to?
But in that case, what happens if you never type COMMIT?
Does it do an auto commit when you exit the program?
What about before that, e.g. if you're running a script
(open script.sql;) does it commit at the end of each script file?
[I have to put setup data from a master database into two separate systems, one running in Oracle 7 on an RS/6000 and the other on a Tandem nonstop SQL server on an S700 with Guardian NSK; both are supposed to conform to SQL standards.]
Regards,
David Wright Harlow UK http://www.northern.co.uk/~dww/
***Boycott Internet Spam*** http://spam.abuse.net/
Received on Sun Aug 29 1999 - 18:25:20 CDT
![]() |
![]() |