In week 15 I told you about the DFDSS program called ADRDSSU. You can use this program to copy existing datasets to new datasets without the tedious work of preallocating the new dataset.

When you make a copy of a dataset using ADRDSSU, you can copy VSAM-datasets as well as NONVSAM datasets. The copy can be succesfully carried out when you have RACF READ access to the dataset you want to copy and there are no enqueues on the dataset (no one must use it while you are copying). The dataset you want to copy to must not exist. Therefore you normally will start with a delete step of the new dataset before the step that makes the copy.

Here is a piece of JCL that copies MY.OWN.DATASET to MY.NEW.DATASET using ADRDSSU:

//ADRDSSU  EXEC PGM=ADRDSSU
//SYSPRINT DD SYSOUT=*
//SYSIN    DD *
    COPY -
    DATASET(INCLUDE(MY.OWN.DATASET)) -
    RENAMEU((MY.OWN.DATASETOUT, -
             MY.NEW.DATASETOUT)) CATALOG
/*

MY.NEW.DATASET will be allocated precisely in the same way as MY.OWN.DATASET except that MY.NEW.DATASET will have only one extent big enough to hold all data from MY.OWN.DATASET exactly as if you had made a move with ADRDSSU. Before you get too excited copying all sorts of dataset, I can tell you that you will find it pretty irritating, that there has to be no enqueues on the dataset, you want to copy. There are also some pitfalls that you must be aware of. ADRDSSU does not copy alternate indexes on a VSAM CLUSTER (because of the RENAMEU parameter) and GDG bases cannot be copied, but GDG's are, which can cause a lot of trouble.