Mainframe, MVS and zOS Discussion

Number three would kill number one and then all 98 survivors would leg it. Or, about half-way around, after a certain amount of frowning and lip-moving, number one would realise he'd made a bad decision in handing on the sword, and he'd leg it, Followed, in a stuttering "peel" by each of t...

It looks like it was written by someone unused to having people understand their programs :-) OK, it is interpreting the code returned from an XML-parsing failure, and that is given as a decimal number represented in a binary field, so always needs to be converted to hex, which is what the code is d...

enrico, you forgot to turn it off again... could be all types of consequences :-)

Validating a file is one thing, validating a program is another. If you have a program A which is testing programs, how do you test program A? It has to work with a infinite number of programs. Now, don't say "you're being silly, it's only working on these strictly-specced programs for a strict...

No, it is the use of what is sometimes called a "Priming Read". If you notice, there are two Reads in the example. One immediately before the loop, and one the last thing in the loop. The code within the loop will only be processed when a record is present. With a Read at the top of the lo...

The last byte of the record will contain the last byte of your data. There is no byte after the last byte of data. There's an end-of-file marker, but you'll never see it.

Use a loop with PERFORM.

Why do you want to do that? It'll take a lot of resources. Why don't you know where what you want is, with a littler more certainty?

What is unclear? If you have multiple record-types of different sizes (or even one record-type, with a varying length) then you can't use RECFM=Fx. Therefore, rather than just give up and go home, we use RECFM=Vx.

There are examples in the documentation. Look up MERGE and the examples there, look up OUTFIL and SECTIONS, remembering to take note of REMOVECC,NODETAIL and TRAILER3. Look at relevant examples.

You see how much easier it is when you post, almost, everything? If you look at your input, you'll see that the key you are sorting on only has data for the first item of a group of records. Same with the account. If you SORT that, you'll get all those spaces together, and added up, as you did and a...

Anuj, since there is no connection between the name SyncSort and the message-prefix WER, and since there are probably thousands of bits of code around the world depending on the prefix WER for extracting information, I doubt that we'll ever see that change :-)

I'd subtract it from zero. You have to tell SORT the size and type of the result. Yes, you can multiply by -1.

If you have the data outside the Mainframe, you can use any scripting language which is available and familiar to you.

I'd use AWK, because I'm more familiar with that.

It does depend on what "edit" means. DITTO can. I should think Filemanager/File-Aid can as well.

I'm tempted to just say No. In fact, since "empty records" are difficult to visualize even, I am going to say, No, you can't create them. Can you show the sysout from running the step so that we can see there error, please? Every time you fail to provide information which is asked, it take...

Can you show the sysout from running the step so that we can see there error, please?

How else would you ever get a decimal-point to appear, if you could never source it from an implied decimal-point?

You will only get spanned records when, as Robert has said, you record-length is larger than the CI size. This will require more than one CI to contain the data,. CI splits are a different thing, nothing to do with spanned records. I've never even considered using spanned records for VSAM, not come ...

Also note the values in the LOCAL-STORAGE SECTION are reset on each entry into a program, if you use LOCAL-STORAGE (this is with or without IS INITIAL).