Re: [Info-ingres] Friday is coming ...
- From: Michael Leo <mleo@xxxxxxxxxxxxxxx>
- Date: Wed, 27 Jul 2005 21:11:05 -0500
At 10:49 AM -0400 7/20/05, Karl & Betty Schendel wrote:
At 9:29 AM -0500 7/20/2005, Michael Leo wrote:Gang,
We usually save the contentious arguments for Fridays ... but maybe we should wait for Roy to come back from vacation before discussing the question a client of mine recently asked:
"Why would you want DDL statements to be transactional?"
Uh, gee ... maybe because the DDL is part of my transaction, i.e. my unit of work, which I want to commit or rollback atomically? Could that be it?
I'd invite the client to explain why DDL should be any different from DML.
-- Karl R. Schendel, Jr. schendel@xxxxxxxxxxxxxx Computer Associates Ingres Development
Two things:
1. I don't disagree with you. I'm just trying to make trouble.
2. As you probably know, Oracle DDL statements, while atomic,
are not transaction friendly. Executing a DDL statement issues
an implicit commit within the Oracle engine. I was pointing out
the the customer how evil this is, but they couldn't understand
why. *Sigh* Just to be clear, in Oracle, this SQL sequence results in a
row actually existing in the wow table and the wow2 table
exists as well. create table wow(neato varchar2(20));
commit;
insert into wow values ('Mike');
create table wow2(neato varchar2(20));
rollback;
select * from wow;
commit;Now that's just wrong.
Cheers,
-- Mikey .
- Prev by Date: RE: [Info-ingres] TERM_INGRES for ssh client.
- Next by Date: Re: [Info-ingres] Friday is coming ...
- Previous by thread: RE: [Info-ingres] Friday is coming ...
- Next by thread: Re: [Info-ingres] Friday is coming ...
- Index(es):
Relevant Pages
|