IBM Mainframe Forum

interesting question!
Does SORT have a simple way to break a sequential file and generate the correct member name based on something (conceptually) like the './ ADD NAME=MEMBERNAME'? I am ready to implement that and disregard any other solution if a sort solution is available.
(PDSLOAD, etc. preserves STAT info, but in my case this is a non-issue!)

ZA

Hello,

I'd suggest manually creating test data for 2 or 3 members/files of different lengths and running sort steps to copy this info into the new pds just to make sure it works.

Sort can generate lots of different output based on the input. For this exercise, is there a value that will always be present in the first record (i.e. a ./ ADD)? If not, how is the member name to be determined?

If you use your sort product, i believe you will need a separate step for each member to be processed - but this should not be a problem as this JCL/control statements can be generated (once the necessary process is determined by the manuall testing).

I thought about (and rejected) the option of a seperate step for each member. I could do that as I am the one who creates the input on a non-mainframe environment, but I do not want to do it. I want to import the whole thing as one file!
I do have a './ ADD NAME=membername' as a separate line before each member. Is there a simple way to split that file and load the library in one SORT step? If not then I will stick with PDSLOAD or my Rexx.
Thanks
ZA

Hello,

I don't know of a way to tell the sort when you want to switch to the next member. . .

Which is why i've used a separate step for each new member.