The characteristics of an indexed sequential data set are established when the data set is created using QISAM. You cannot change them without reorganizing the data set. The DCB parameters that establish these characteristics are: BLKSIZE, CYLOFL, KEYLEN, LRECL, NTM, OPTCD, RECFM, and RKP.

A data set processed with QISAM can have unblocked fixed-length records (F), blocked fixed-length records (FB), unblocked variable-length records (V), or blocked variable-length records (VB).

QISAM can create an indexed sequential data set (QISAM, load mode), add additional data records at the end of the existing data set (QISAM, resume load mode), update a record in place, or retrieve records sequentially (QISAM, scan mode).

For an indexed sequential data set, you can allocate space on the same or separate volumes for the data set's prime area, overflow area, and cylinder/master index or indexes. For more information about space allocation, see z/OS MVS JCL User's Guide.

QISAM automatically generates a track index for each cylinder in the data set and one cylinder index for the entire data set. Specify the DCB parameters NTM and OPTCD to show that the data set requires a master index. QISAM creates and maintains as many as three levels of master indexes.

You can purge records by specifying the OPTCD=L DCB option when you allocate an indexed sequential data set. The OPTCD=L option flags the records you want to purge with a X'FF' in the first data byte of a fixed-length record or the fifth byte of a variable-length record. QISAM ignores these flagged records during sequential retrieval.

You can get reorganization statistics by specifying the OPTCD=R DCB option when an indexed sequential data set is allocated. The application program uses these statistics to determine the status of the data set's overflow areas.

When you allocate an indexed sequential data set, you must write the records in ascending key order.