For nonrelational sources, such as IMS databases, VSAM files, sequential data sets, and flat files, data maps are required to perform bulk data movement operations, and to provide the metadata used to define capture registrations and extraction maps for CDC.

For DB2 sources, you can also define data maps so that you can split the data in a column that contains multiple fields into separate fields.

A data map for a flat file or sequential data set must be associated with a data file that defines the record types and the data in the data map.

When you add a data map, optionally import a copybook to define the layout of the data in a data source and, for an Adabas data map, import an Adabas field definition table (FDT) and Adabas keys. Alternatively, you can define the layout of the data by manually defining one or more records and tables in a data map.

Records in data maps define the field layout of the records in a source or target. Tables in data maps provide a relational view of the data. PowerExchange requires a relational view to access data sources by using the SQL language, as follows:

PowerExchange maps fields in data maps to the byte level. Some copybooks, such as PL/I copybooks, have bit-level mappings. If you import a copybook that has bit-level mappings into a data map, PowerExchange concatenates the bit fields into a byte field, naming the field by using the field name of the first bit field. In a data map, you can define a user-defined field that uses one or more PowerExchange functions in an expression to separate this field back into bits. Then, in PowerCenter, use an Expression transformation to calculate the numeric value of the combined bits in the bit string.

You can run a database row test on a data map to verify that it can retrieve source data.

For information about opening, closing, and deleting data maps, and searching for fields in records and columns in tables in data maps, see PowerExchange Resources.