A clustering index determines the physical order in which data rows are stored on disk. When a clustering index is defined, DB2 attempts to maintain the data sequence matching the index key order. Other index types (unique, partitioning, secondary) only affect logical access paths.