The SPARSEINDEX option controls the type of index algorithm that composites use to load and access their values. The value of SPARSEINDEX at the time a named composite is defined, or an unnamed composite is created, determines the type of algorithm the composite uses by default. When you specify an index algorithm in a DEFINE COMPOSITE statement, this overrides the default specified by the SPARSEINDEX option.

Choosing an index algorithm is important only in regard to performance issues. Any recommendations are for the version of Oracle OLAP that is associated with this documentation. You can test how using different algorithms affect performance by using a CHGDFN statement to change the algorithm for a composite (for example, before loading data).

Data Type






A standard indexing method that is recommended for composites. Use BTREE unless you are an advanced user. BTREE tends to group similar values together, which results in better locality of access. BTREE is the default algorithm.


A standard indexing method that should only be used when a composite has only two or three base dimensions. HASH is generally not recommended for composites since using HASH results in a very large index table, which can be too large to fit into memory.


Example 6-100 Using the HASH Algorithm

The following example sets SPARSEINDEX to HASH so that composites that are subsequently defined or created will use the HASH index algorithm by default.