Unix comparable function to mainframe

[Larry Scheetz]

Admins, Our Mainframe Admin wants to know if I have a function comparable in HPUX to what they call "Generation Data Sets" for file handling. I am not familiar with this or seen anything like this before. Does anyone know of anything comparable in Unix?
Thx

[A. Clay Stephenson · Accepted solution]

That is a foreign concept to UNIX. Essentially what he is talking about is a set of closely-related files which can be given version numbers. Consider a parents-children-pets database. At some point in time, we might have parents;0, children;0, and pets;0. At some other point in time we might have parents;1, children;1, and pets;1. The ";N"'s represent the version and the application simply sees parents, children, and pets rather than the filename;version.

In UNIX any of this versioning is considered to be an application task rather than an OS built-in.

[Steven Schweda]

A quick Google search for
Generation Data Sets
found:
http://publib.boulder.ibm.com/infocenter/zoslnctr/v1r7/topic/com.ibm.zconcepts.doc/zconcepts_175.html
which offers a brief explanation.

Note that ";nnn" version notation is used in
VMS, not in "mainframe" systems.

[TwoProc]

It's a nice feature that I wish both Unix and Windows had. I've not seen it since I used things Vaxen.

[A. Clay Stephenson]

Note that the filename;N notation was simply used as an example. The exact details of how the OS does this versioning of related sets of files is implementation dependent but the concept remains the same. It's possible, in principle, to emulate this behavior but it would very much depend upon how the application were coded. For example, if my previous example database always looked in /mydir for parents, children, and pets then forget it. If it were configured to use an environment variable then it would be possible to use ${DBDIR}/parents, ${DBDIR}/children, and ${DBDIR}/pets where ${DBDIR} could house different copies of the data. Again, unlike with some OS'es, it would be your task to handle the copying and segregation of the data under UNIX.