Comments on: Did You Know That About PL/SQL Variables?

By: Tim Hall

Tim Hall — Tue, 05 Jun 2007 08:04:59 +0000

Regarding Documentation:

Yes, the bug I raised was against the documentation. It originally quoted the old 2000 byte limit.

I've not checked recently, but the old 10.1 documentation was left with the incorrect value.

Cheers

Tim...

By: Jonathan Lewis

Jonathan Lewis — Sat, 02 Jun 2007 15:25:23 +0000

There is a drawback to this approach. If the actual usage varies significantly in the lifetime of your session, then dynamic allocation can result in multiple child cursors for the same statement.

On a highly concurrent system, this may be something you would prefer to avoid, since all three versions will be held on the same library cache latch, and multiple versions means longer hold times.

There is a short note about this type of thing, with some follow-up discussion, at

http://jonathanlewis.wordpress.com/2007/01/05/bind-variables/

By: Lia Green

Lia Green — Sat, 02 Jun 2007 02:46:15 +0000

I have a question…are people still interested in getting certified? Cerebral Solutions is offering a 6 day certification and can’t fill classes. It’s unbelievable. Has certification lost its shine?

By: Eddie Awad

Eddie Awad — Fri, 01 Jun 2007 16:23:15 +0000

And here is an excerpt from the PL/SQL User's Guide and Reference 10g Release 2: You might need to allocate large VARCHAR2 variables when you are not sure how big an expression result will be. You can actually conserve memory by declaring VARCHAR2 variables with large sizes, such as 32000, rather than estimating just a little on the high side, such as by specifying 256 or 1000. PL/SQL has an optimization that makes it easy to avoid overflow problems and still conserve memory. Specify a size of more than 4000 characters for the VARCHAR2 variable; PL/SQL waits until you assign the variable, then only allocates as much storage as needed.

By: Tim Hall

Tim Hall — Fri, 01 Jun 2007 07:11:28 +0000

The limit is actually >4000 in 10g. I know because I spotted it and raised it as a bug (4330467) on metalink, only to be told it was a "feature" of 10g.

Cheers

Tim...

By: joel garry

joel garry — Fri, 01 Jun 2007 00:05:25 +0000

Shouldn’t that be “… once you go above 1999 characters?”

Sounds like a recipe for unnecessary performance issues to me. Worry less about the few M of your users address space and more about many cycles of your cpu, since the latter will suddenly enhance any potential latching issues, but when do you have to search any wasted user space?