There are three formats that the unload data set can take. The
primary difference between them is which records can appear in the
data set. These formats are:
- Invalid Format
All records after the COPYR1 (if any) are undefined.
The condition occurs when an unload operation is ended because of
an error. The COPYR1 is re-written as the first record in the container
data set with "Invalid Format" as part of error clean-up.
- Old Format (Pre-PDSE)
- There may be a note list record for each member.
- There are no attribute records.
- The original data set was a partitioned data set, not a PDSE.
- All DASD addresses are valid for a real device, and the DEB and
DEVTAB information comes from the DASD device which held the original
data set.
- The second batch of data set label information (starting 46 bytes
into COPYR1 record) is not present.
- New Format
- Note list records are now used by the linkage editor and other
applications to record relocatable addresses of records inside the
member.
- Attribute records may be present if the original data set is a
PDSE.
- Records from a PDSE contain DASD addresses from an artificial
device that has 256 tracks and 65536 cylinders and tracks of 16M bytes.
This convenience maps the maximum number of possible PDSE RLTs and
MLTs according to the restrictions for accessing PDSEs with BPAM documented
in the z/OS DFSMS Using Data Sets.
While these addresses of consecutive
records are strictly ascending, some addresses are not to be used.
Record numbers for PDSE members are always odd starting with 1 and
continuing 3, 5, 7, 9... Even record numbers are reserved.
- The second batch of data set label information (starting 46 bytes
into the COPYR1 record) is valid.