Teradata Warehouse Miner requires data sources to be defined in a certain manner. Either File, User, System or “Simple” (IP Address or Host Name) data sources can be used. A connection is only required when selecting data from the user interface, when an analysis executes or when result tables are displayed.
When a data source is defined, the name of the Default Database is used as the Teradata Warehouse Miner Source Database. This is the Teradata database where the tables to be analyzed reside. This is only a default and can be changed on the Connection Properties dialog as described in Using Teradata Warehouse Miner.
- The Metadata Database is the Teradata database containing the Teradata Warehouse Miner metadata.
- The Result Database will contain any result tables or views.
- The Statistical Test Database contains the metadata tables that support Statistical Test functions.
- The Publish Database contains the metadata tables that hold “published” analyses.
- The Advertise Database contains the metadata tables that hold information about “advertised” output tables, views and procedures.
Once these Teradata Warehouse Miner databases are changed, this information is saved on a per Data Source basis. When you reconnect utilizing a particular data source, Teradata Warehouse Miner remembers the Source, Result and Metadata database settings from the prior connection.
Depending on the version of Teradata and the version of the ODBC driver for Teradata being used, the session character set requested on an ODBC data source may need to be set. See the Orange Book publication 541-0004068-C02 Getting Started: International Character Sets and the Teradata Database for more information.