A VSAM index can consist of more than one index level. Each level contains a set of records with entries giving the location of the records in the next lower level.

Figure 1 shows the levels of a prime index and shows the relationship between sequence set index records and control areas. The sequence set shows both the horizontal pointers used for sequential processing and the vertical pointers to the data set. Although the values of the keys are actually compressed in the index, the figure shows the full key values.

Figure 1. Levels of a Prime Index

Sequence Set. The index records at the lowest level are the sequence set. There is one index sequence set level record for each control area in the data set. This sequence set record gives the location of data control intervals. An entry in a sequence set record consists of the highest possible key in a control interval of the data component, paired with a pointer to that control interval.

Index Set. If there is more than one sequence set level record, VSAM automatically builds another index level. Each entry in the second level index record points to one sequence set record. The records in all levels of the index above the sequence set are called the index set. An entry in an index set record consists of the highest possible key in an index record in the next lower level, and a pointer to the beginning of that index record. The highest level of the index always contains only a single record.

When you access records sequentially, VSAM refers only to the sequence set. It uses a horizontal pointer to get from one sequence set record to the next record in collating sequence. When you access records directly (not sequentially), VSAM follows vertical pointers from the highest level of the index down to the sequence set to find vertical pointers to data.