z/OS: Efficient direct access to a multi-volume data set

I have a multi-volume MVS data set A.B.C with parameters PO=PS and RECFM=FS. The code below runs slowly (Java is used for illustration, in C it will be fseeko call) :

ZFile zfile = new ZFile(//A.B.C, rb,type=record);
file.seek(1000,0);
file.seek(2000,0);
file.seek(4000,0);

Run time for each seek: 1-2 seconds.

However, the same code runs fast for a single volume data set. Run time for each seek: 5-10 milliseconds.

I need Java code that performs fast (comparable to single volume) direct access to a multivolume data.

Acceptable solutions:

Implementation of a Java method, C, or assembler function that efficiently positions (in milliseconds) file pointer to a given record in both multivolume and single volume data set.

For clarity, Java prototype is:

void fastseek (ZFIle file, long recordNumber);

Note: C/Assembler solution should work in 64-bit architecture.
Note: Expected file size up to 20 Gb,
About the recuiter
Member since Nov 11, 2022
Manoj Kumar
from Stredochesky, Czech Republic

---

---

Looking for help? Checkout our video tutorial
How to search and apply for jobs

How to apply? Do you have more questions about the Job?
See frequently asked questions