IBM Mainframe Forum

How to copy a file without changing the creation date?

IEBGENER, IEBCOPY, ADRDSSU, ...

Pretty much anything will copy a data set without changing the creation date of the source file.

Mr. Sample is correct. You only change the creation date of a dataset by deleting and reallocating the dataset.

Do not use the term "file" when referring to a mainframe dataset. A mainframe "file" exists only on magnetic tape.

Hello and welcome to the forum,

How to copy a file without changing the creation date?

Does this imply that you experienced the creation date changing?

As mentioned, the creation date of the input file does not change. . .

Is someone trying to hide that there were later modificatons to the data? This would be a reason for termination in some organizations. . .

I read this topic as not changing the creation date of the NEW dataset from what the original was.

I read this topic as not changing the creation date of the NEW dataset from what the original was.

in which case copy it on the day it was created!

Hi Spock,

I read this topic as not changing the creation date of the NEW dataset from what the original was.

Yup, which is a problem as the organization needs to know the file was modified later. . . (At least places i've supported do).

Might do more than just raise a few eyebrows during an audit as well. . .

d

After thinking about this for a while, I realized the real question might be, "Can I falsify the creation date of a dataset?" The answer is there is a partially documented method to allocate a dataset with a "false" creation date. This method actually goes back to OS/360 and was used by IEHMOVE and is now also used by DFHSM when it restores a dataset from migration or backup. Once the dataset has been allocated there is no method to alter its creation date.

If the real question is, "Can I obtain the modification date (and perhaps time, too)?", the answer is this is not a dataset attribute stored with the dataset, so you can get it from analyzing SMF data, but not from analyzing the VTOC. You do have DS1DSCHA in DS1DSIND to indicate a dataset has been altered, but that's far from a date.