Re: Get ref cursor data dynamically - HOW ????

From: Jim Kennedy <kennedy-down_with_spammers_at_attbi.com>
Date: Sat, 15 Feb 2003 00:37:43 GMT
Message-ID: <rZf3a.103214$SD6.5545_at_sccrnsc03>


Remember if you dynamically create a table then it issues a commit and you can't roll it back. eg
insert into table_blah(1);
create temp_table (number not null);
rollback;
if you then do
select * from table_blah;
You will get
1
(assuming nothing else in the table)
Jim

--
Replace part of the email address: kennedy-down_with_spammers_at_attbi.com
with family.  Remove the negative part, keep the minus sign.  You can figure
it out.
"V. Pryadkin" <vpryadkin_at_yahoo.com> wrote in message
news:33c9717f.0302141517.60754848_at_posting.google.com...

> Jim,
>
> I totally agree with you, and what is more I am pretty sure that any
> dynamical things must be avoided whenever it's possible, especially
> dynamic DDL's. So I'd say my strange topic is rather an indicator of
> poor "dynamic" experience, but not poor programming bents :o).
>
> If it's interesting I can explain what I really want, probably there
> are known solutions which I just don't know or you can just give me
> good advice. The issue is called "regression test". Shortly that means
> "run pl/sql stuff -> save the results somehow -> rollback the database
> to the initial state -> run same stuff again -> match out what you've
> got now with what you saved during the first pass". In fact results
> mean three things:
> 1) output variables
> 2) ref cursors
> 3) changes in database tables
>
> Sure, within the session I can use global temporary tables, nested
> tables etc as a storage. Actually there is no problem with (1) and
> (2), there are almost no dynamic nor temporary things to do. But (3)
> means a bunch of various record structures, so honestly I haven't
> decided yet how to "save a ref cursor twice & compare" easily. But at
> least I need to manage ref cursors dynamically, there are no
> questions. Let's say QA guys will use my testing API's, but definitely
> they are not supposed to be very good in pl/sql to create their
> testing scripts, especially hardcoding record types for each given ref
> cursor or something like that.
>
> Steve.
>
> "Jim Kennedy" <kennedy-down_with_spammers_at_attbi.com> wrote in message
news:<Wh83a.97484$2H6.1439_at_sccrnsc04>...
> > In addition to Brian's observation I strongly recommend you don't
> > dynamically create temporary tables. This is usually an indication of
poor
> > programming practice. (read crutch from other databases where it is a
> > necessity) It will lead to piss poor performance and scalability. If
you
> > really really really just gotta have them then use global temporary
tables
> > (see docs). The use of inline views etc. makes it highly unlikely that
you
> > need to create temporary tables on the fly. (see asktom.oracle.com for
> > discussions and other options)
> > Jim
Received on Sat Feb 15 2003 - 01:37:43 CET

Original text of this message