Records are loaded into a key-sequenced data set (KSDS) in key sequence, the order defined by the collating sequence of the contents of the key field in each of the records. Each record has a unique value in the key field, such as employee number or invoice number. VSAM uses an index and optional free space to insert a new record into the data set in key sequence.

A key-sequenced data set is prepared for restart by the restoration of any statistical information (such as number of records inserted) to its checkpoint status.

For a VSAM KSDS, restart does not erase any data except in create mode. It does, however, detect modification of the data set by either the checkpointed program or another program that used the data set between the checkpoint and the restart. If the data set is modified, the restart is terminated, unless you override the testing of the data set by using the AMP=‘CROPS=NCK’ subparameter in the DD statement for the data set.