The Relational MR Database (RDB) CDSC reads and writes case data to a relational SQL Server database.

The current version of this DSC reads and writes the data through OLE DB and is known as RDB DSC 2 to distinguish it from the older ODBC-based version of the DSC, which was known simply as RDB DSC. The RDB DSC 2 has an internal name of mrRdbDsc2.

Although the RDB DSC is no longer included with the UNICOM Intelligence Data Model, the RDB DSC 2 supports older databases that use the original RDB DSC schema. The RDB DSC hierarchical data schema (introduced in version 6) is created when unexpanded levels appear in the metadata. For more information, see Relational MR Database CDSC schema.

The RDB DSC 2 and RDB DSC Relational MR Database CDSC schema both use several tables to store case data, which gives many advantages over a single-table schema, in which each respondent corresponds to a single record and the responses are stored in the table columns.

Modern database software is designed to use relational schemas and a single table schema has many disadvantages. When representing case data from a long questionnaire, a single table schema can easily exceed the byte and column restrictions placed on a row. This means that the schema can overrun into multiple tables, which in turn raises access issues, because all users must be granted permission to create tables. A single table schema contains a lot of empty space, because it must be set up to contain all possible data permutations, rather than simply allocating space for each response as it is made. Finally, writing data requires that data records be updated, which means that the record must first be found, creating unnecessary overheads that can lead to performance bottlenecks during data collection.

The schema addresses these issues as follows: