In a key-sequenced data set, there is an index entry for each control interval that contains a segment of a spanned record. All the index entries for a spanned record are grouped together in the same section. They are ordered from right to left according to the sequence of segments (first, second, third, and so on).

Only the last (leftmost) index entry for a spanned record contains the key of the record. The key is compressed according to the rules described above. All the other index entries for the record look like this: