Analytics Database is composed of parallel units known as “AMPs." Each node in the configuration has a predefined number of AMPs, and typically the rows of a table are spread across all AMPs in the system.
Maps (Hash Maps) determine how to distribute data from individual table rows among the AMPs. Maps list one or more AMPs. Every table is associated with a map to distribute the table data (rows) among the AMPs in the map.
Other objects such as user defined functions (UDF) are also associated with maps. A UDF is placed on the AMP that stores the associated table's rows.
- Contiguous maps include all AMPs within a specified range. Analytics Databasee creates contiguous maps during a system initialization or reconfiguration.
- Sparse maps include a subset of AMPs from a contiguous map. Analytics Database provides two sparse maps: one sparse map for a subset of the AMPS and one sparse map for one AMP.
MAPS Manager selects the optimal map for each object and provides the recommendation in System Monitoring or lists the target map in Object Workflows.
The order of map selection is first the system default map and the associated system created sparse maps, then any user created map that the user has been granted permissions to use. For System Monitoring, the credentials entered in the settings determine the maps for use. For Object Workflows, the logon credentials determine the maps that are selected or available. The user created maps must have a unique number of AMPs. If there are two user created maps with the same number of AMPs, MAPS Manager selects the map that alphabetically comes first. For example, the map Ann03 is selected instead of the map Vin03.
You can change the target map after analysis using workflow options. For more information, see Teradata® Database Design.
For the maps defined for a system, see the Maps tab. You can also use this view to create new sparse maps and grant or revoke permissions for users and roles.