Re: Reset data

Arnold,
There USED to be performance problems with IBM's INITIALIZE code, but I
thought they fixed most (all?) of them a while ago. Have you run tests with
"recent" versions of Enterprise COBOL that still show performance problems?

--
Bill Klein
wmklein <at> ix.netcom.com
"Arnold Trembley" <arnold.trembley@xxxxxxxxxxxxxxxx> wrote in message
news:BC6Fh.19799$as2.5112@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx


Pete Dashwood wrote:

"Howard Brazee" <howard@xxxxxxxxxx> wrote in message
news:chv5u214c3r2strckf1h2r8msh9pv42ne5@xxxxxxxxxx
(snip)
One thing I sometimes do in a program with such a table (especially if
some of the variables are packed or binary), is to create an
initialization table. If I need to reinitialize the table, I just
MOVE ZERO-TABLE to TABELLE.


Again, something I don't understand.

How would the result of doing that (taking twice as much storage as you
actually need, and running the risk of forgetting to update one of the
tables when the structure changes) differ from...

INITIALIZE TABELLE ?

Pete.

If you are INITIALIZE'ing a structure, the generated code will have an
explicit move for each elementary field. I know it works this way in
Enterprise COBOL for zOS. If you INITIALIZE the same structure in a loop many
times, it can create some performance delays for certain applications. I will
grant you that this kind of optimization is not typically needed in online
applications, but it could be a performance bottleneck in some cases.

With kindest regards,


--
http://arnold.trembley.home.att.net/



.