International Business Machines Corporation
ENHANCING SPARSE INDEXES

Abstract:

A data structure associated with a sparse index is determined to include a plurality of redundant keys with at least one set of duplicate keys. The at least one set of duplicate keys is ranked, according to a set of criteria. According to the ranking, a first set of duplicate keys from the at least one set is selected. In place of the first set, a first guard node is inserted. The first guard node includes a first key value identical to the first set of duplicate keys and is linked to a first set of field nodes representing a first set of field values associated with the first set of duplicate keys.