Bookshelf Home | Contents | Index | Search | PDF |
Siebel Marketing Guide > External Data Mapping >
Understanding Joins
Siebel Marketing can retrieve data from multiple RDBMS tables. To extract and process this data, the Marketing Server must be able to associate and merge related data from these data sources into Siebel Repository tables and fields.
Joins allow Siebel Marketing to match and combine records. Siebel Marketing joins work like those in any true relational database. Records from two tables are merged when join keys match in each table. Although tables in your RDBMS have defined joins, you also need to define joins in Siebel Marketing. Siebel Marketing uses its own join definitions because the Marketing Server can join tables across separate RDBMS data sources that do not share the same physical database.
The Marketing Server decides which of three join methods to use to retrieve data from mapped tables. The join method selected depends on whether the join fields are also mapped to the customer hierarchy. The Siebel Marketing Server applies rules in the following order to determine the method it will use to retrieve data. Usually, merge joins and cache joins provide much better performance than nested loop joins.
- Merge. When customer hierarchy mappings are consistent with joined fields, the server uses the merge method, even if the cache flag is turned on.
For example, when you have two tables (Table1 and Table2), the Marketing Server will use the join method if the following conditions are satisfied:
- The join key value from Table1 matches the join key value from Table2.
- Both the join key value from Table1 and the join key value from Table2 are in the customer hierarchy.
If one of the fields is not in the customer hierarchy, the Marketing Server will not use the Merge method.
- Cache. If the Merge rule is not true and you turn on the cache flag for the join, the server will cache (read into memory) the child table.
- Nested loop. If the cache rule is not true, the server uses the nested loop method. This method is only used when Merge and Cache do not apply. The nested loop method should be avoided to prevent performance slow downs, especially for large tables or when joining tables without indexes.
For more information about joins, see the following topics:
Bookshelf Home | Contents | Index | Search | PDF |
Siebel Marketing Guide, Version 7.5, Rev. A Published: 18 April 2003 |