z/OS DFSMS Using Data Sets
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF

Spanned Format-VS Records (Sequential Access Method)

z/OS DFSMS Using Data Sets

Figure 1 shows how the spanning feature of the queued and basic sequential access methods lets you create and process variable-length logical records that are larger than one physical block. It also lets you pack blocks with variable-length records by splitting the records into segments so that they can be written into more than one block.

Figure 1. Spanned Format-VS Records (Sequential Access Method)

The format of the BDW is as described in Figure 1.

When spanning is specified for blocked records, QSAM attempts to fill all blocks. For unblocked records, a record larger than the block size is split and written in two or more blocks. If your program is not using the large block interface, each block contains only one record or record segment. Thus, the block size can be set to the best block size for a given device or processing situation. It is not restricted by the maximum record length of a data set. A record can, therefore, span several blocks, and can even span volumes.

Spanned record blocks can have extended BDWs. See Block Descriptor Word (BDW).

When you use unit record devices with spanned records, the system assumes that it is processing unblocked records and that the block size must be equivalent to the length of one print line or one card. The system writes records that span blocks one segment at a time.

Spanned variable-length records cannot be specified for a SYSIN data set.

Go to the previous page Go to the next page

Copyright IBM Corporation 1990, 2014