A non-unique secondary index can be value ordered which means the non unique secondary index can be sorted on the key values themselves rather than on the corresponding hash codes. This is useful for range queries where only a portion of the index subtable will be accessed. With a value-ordered non unique secondary index, only those blocks in the non unique secondary index subtable that are within the range are scanned. It must be a number value, up to 4 bytes, versus a longer character column. DATE is the most commonly used data type. The actual data value is stored as part of the non unique secondary index structure.