Mainframe, MVS and zOS Discussion

We have 1 million records in a file. One of the records has bad data. To solve the abend because of this bad record, I've brought the file with all the records to my test environment. Once the file is successfully transferred, I delete that bad data record. Now can I refer this updated file to job running in the production environment?

I want to know whether we can override data of production environment with data of test environment, as described above. I did not know the answer so I am asking here. If someone an help.

Overriding the data in Production should be achievable.

Is this REALLY an Interview Question?

PS.: Sounds like the interviewer had a problem in hand and put forward it as an interview question. Perhaps from someone like who is juggling to know - love what you do or do what you love!

To original question - you can do it but much depend on the site policies.

The issue will be the security authorisation. In general, if things are set up well, then No. But it all depends on how a site is set up.

If you can copy the data to your test environment, that is wrong. That should require authorisation, and someone to do that for you, the same on the way back (to a dataset-name that suits Production).

If you are a specific Support person, some of those things will be different than if you are a Developer.

The only answer that can be given in general is No, you can't mess about with Production data yourself. Which means you can't be the one to forget to back it up, you can't be the one to accidentally erase the data before it is backed-up, accidentally change all the records instead of one, etc.

Oh, and you can't do it deliberately either, at least not without others knowing, and having convinced those people of why you need access to the data.

even if can be done from a technical point of view
it is a bad practice ...

run the fixup process in the production environment.
naturally with the appropriate audit trail

I'll take a stance combining that of Mr. Collins and that of Dr. Sorichetti.

You should not be able to merely copy data sets from production to test; you should be able to do so only with the proper tools to provide for proper masking of the data. The fix-up should be developed in the test environment, but be executed in the production environment; you should never be able to copy data from test to production[sup]1[/sup]. And, of course, the fix must be developed by a competent analyst, not a fresher who lied about knowing Rexx.

[sup]1[/sup]I concede that in a von Neumann architecture the line between "data" and "program" can get a little fuzzy; nonetheless, in the real world that line is clear enough.