REXX to find /* Define and write the next 9 lines

markzasz · New User Joined: 04 Apr 2017 Posts: 9 Location: USA

Hello,

I have a datasets that contains 22 millions lines and 450,000 IDCMAS Define statements along with many other commands. I would like to scan the dataset and create multiple output files that contain the Define statements based on the usercatalog.

my defines look like this in the dataset.

/* DEFINE -
NONVSAM( -
NAME(PCMSS.MS3AT.EXTRACT.INDX.BASE.G0001V00) -
RECATALOG -
VOLUMES(MG4W12) -
DEVICETYPES(3390) -
) -
CATALOG(CATALOG.DZOSPLX.PROD05);

So when you find "/* DEFINE"
- save DEFINE and the next 7 lines
- and write the saved info to one of five output files.

the reason for five output files is we have 5 User Catalogs that these defines are going to and we would like to break the defines up by User Cataloag.

Thank you,

Mark...

Joerg.Findeisen · Posted: Fri Jun 11, 2021 4:30 pm

Consider using a SORT product for this mass data processing, more fun.

Pete Wilson · Posted: Fri Jun 11, 2021 6:48 pm

This seems a very odd thing to do unless perhaps you're cloning the volumes and making them available on other Sysplexes or VM Z/OS guests or something.

Trying to catalog the 'same' physical dataset, as implied by volser being coded, in 4 different catalogs on the same Sysplex? I'm not even sure it would be possible if they're SMS managed because the catalog backward pointer in the datasets NVR record in the volumes VVDS can only point to one catalog, which would get a mismatch during RECATALOG processing if a different catalog was specified.

markzasz · New User Joined: 04 Apr 2017 Posts: 9 Location: USA

I am not cloning the same dataset, i have a dataset with 25 million lines of IDCMAS statements created by a TACM Program, in this dataset i have 492,000 IDCAMS Define statements i would like to extract and place in a separate dataset that only contains the IDCAMS Define statements.

So one dataset with 492,000 of these.....

DEFINE -
NONVSAM( -
NAME(PCMSS.MS3AT.EXTRACT.INDX.BASE.G0001V00) -
RECATALOG -
VOLUMES(MG4W12) -
DEVICETYPES(3390) -
) -
CATALOG(CATALOG.DZOSPLX.PROD05);

jerryte · Posted: Fri Jun 11, 2021 7:23 pm

Try using SuperSearch (ispf option 3.15). Set process option LPSF which will list lines before and after a find. Then define and edit a Statements DSN and code a "LPSFV 7" to specify lines instead of the default 6. Using the results from this you can then process with a Rexx.
NOTE: the EXECIO cannot read a large quantity of lines. You will need to put inside a loop and process maybe 1000 lines at a time. Good luck.

Rohit Umarjikar · Posted: Fri Jun 11, 2021 7:26 pm

Mark Zaszczurynski, Please don't post twice the same question under different section.
ibmmainframes.com/viewtopic.php?p=352327#352327