To create an indexed sequential data set by the one-step method,
take the following actions:
- Code DSORG=IS or DSORG=ISU and MACRF=PM or MACRF=PL in the DCB
macro.
- Specify the following attributes in the DD statement:
- DCB attributes DSORG=IS or DSORG=ISU
- Record length (LRECL)
- Block size (BLKSIZE)
- Record format (RECFM)
- Key length (KEYLEN)
- Relative key position (RKP)
- Options required (OPTCD)
- Cylinder overflow (CYLOFL)
- Number of tracks for a master index (NTM)
- Space requirements (SPACE)
- To reuse previously allocated space, omit the SPACE parameter
and code DISP=(OLD, KEEP)
- Open the data set for output.
- Use the PUT macro to place all the records or blocks on the direct
access volume.
- Close the data set.
The records that comprise a newly created data set must be presented
for writing in ascending order by key. You can merge two or more input
data sets. If you want a data set with no records (a null data set),
you must write at least one record when you allocate the data set.
You can subsequently delete this record to achieve the null data set.
Recommendations:
- If you unload a data set so that it deletes all existing records
in an ISAM data set, at least one record must be written on the subsequent
load. If no record is written, the data set will be unusable.
- If the records are blocked, do not write a record with a hexadecimal
value of FF and a key of hexadecimal value FF. This value of FF is
used for padding. If it occurs as the last record of a block, the
record cannot be retrieved. If the record is moved to the overflow
area, the record is lost.
- After an indexed sequential data set has been allocated, you cannot
change its cms characteristics. However, for
added flexibility, the system lets you retrieve records by using either
the queued access technique with simple buffering or the basic access
method with dynamic buffering.