IBM Mainframe Forum

Hi,

I want to edit a jcl in which i have to replace a dataset with another in more than one place.

can i do it in one go rather than copying the new one in each place?

Hello,

If you are using the tso/ispf editor (option 2 from the primary menu or selecting from the dataset list in 3.4)
use the command --> c 'dsn1' 'dsn2' all. If you are using some other editor, you need to explain which one you are using.

Much can be learned from the online tutorial.

hello dick,

This is working fine for me.i tested with small ds name.But our site standards suggests to write dataset name like this DEVL.o82201.CREDP100.TEMP.UNBNO.so i am unable to include all the command in the required space.so is there any way to increase the available space in cmd line or anyother way...

(i am asking this because i am working as a production control analyst and need to fix the failed batch jobs by replacing the datasets.)

please help.

nn_az wrote:... But our site standards suggests to write dataset name like this DEVL.o82201.CREDP100.TEMP.UNBNO.

There are two problems with your "site standard:' You cannot use lower case characters in a dataset name, and the last character in a dataset name cannot be a period(.).

There are three possible solutions to the problem you stated.

• You probably only need to change a small and most likely unique part of the dataset name. Just change that part.
• Change the long string to a short, hopefully unique, short string, then change the short string to your final name.
• This is a desperation solution, but it has worked for me in the past: use the line mode TSO EDIT command. There are no practical length restrictions in the source and new test strings in its change commad.

knowing the tools of the trade is the first thing to do...
f 'somelongstring'
c * 'someotherlongstring' all

the command line will hold a 44 char dsn for the find,
unfortunately the CHANGE ALL will hold only 40 chars

but You can split the change in two
f 'first.part.'
c * 'first.changed.' all
f '.second.part'
c * '.second.changed' all

the topic is not JCL related, might be moved to the ISPF section
and the title changed to ....
how to issue commands involving long operands

Something else you might find useful is to exclude all the lines which don't meet your search criteria (you can even get the notification of the existence of an excluded line). Then you only need to do a change for the part of the datasetname that is different, but can do so in confidence that it is not going to change anything by accident.

many thanks to all experts.

Also, when if do a find for your datasetname-to-be-changed you could then do



* tells the editor to use the previous find text (from either find or change).

If you don't put "all" on the end of the change but want to do them one at a time, you can


for each line. Look at RFIND and RCHANGE as well. Useful on function keys if not there already.

Or if you do one change correctly and want then to do all the rest, you can



Full procedure might be



It is well worth checking all of this out in the online help/tutoria for the editor. Find and Change are very powerful, if you are aware of what you can do with them.

Edit: I keep forgetting to add that you could also make a little editor macro if this is a common task for you.

Why the c2x'es? To prevent problems with embedded '&'s and '"'s...

prino wrote:
[...]

Why the c2x'es? To prevent problems with embedded '&'s and '"'s...

Very nice tip Prino. Thanks, I'll remember it.