Records can be retrieved sequentially using keyed access or addressed access.

Keyed Sequential Retrieval. The first time your program accesses a data set for keyed sequential access (RPL OPTCD=(KEY,SEQ)), VSAM is positioned at the first record in the data set in key sequence if and only if the following is true:

1. Nonshared resources are being used.
2. There have not been any previous requests against the file.

If VSAM picks a string that has been used previously this implicit positioning does not occur. Therefore, with concurrent or multiple RPL's, it is best to initiate your own POINTs and positioning to prevent logic errors.

With shared resources, you must always use a POINT macro to establish position. A GET macro can then retrieve the record. Certain direct requests can also hold position. See Table 1 for details on when positioning is retained or released. VSAM checks positioning when processing modes are changed between requests.

For keyed sequential retrieval of a fixed-length or variable-length RRDS, the relative record number is treated as a full key. If a deleted record is found during sequential retrieval, it is skipped and the next record is retrieved. The relative record number of the retrieved record is returned in the argument field of the RPL.

Addressed Sequential Retrieval. Retrieval by address is identical to retrieval by key, except the search argument is a RBA, which must be matched to the RBA of a record in the data set. When a processing program opens a data set with nonshared resources for addressed access, VSAM is positioned at the record with RBA of zero to begin addressed sequential processing. A sequential GET request causes VSAM to retrieve the data record at which it is positioned, and positions VSAM at the next record. The address specified for a GET or a POINT must correspond to the beginning of a data record; otherwise the request is not valid. Spanned records stored in a key-sequenced data set cannot be retrieved using addressed retrieval.

You cannot predict the RBAs of compressed records.

GET-previous (backward-sequential) processing is a variation of normal keyed or addressed-sequential processing. Instead of retrieving the next record in ascending sequence (relative to current positioning in the data set), GET-previous processing retrieves the next record in descending sequence. To process records in descending sequence, specify BWD in the RPL OPTCD parameter. Select GET-previous processing for POINT, GET, PUT (update only), and ERASE operations. The initial positioning by POINT, other than POINT LRD, requires that you specify a key. The following GET-previous processing does not need any specified key to retrieve the next record in descending sequence.

GET-previous processing is not permitted with control interval or skip-sequential processing.