Dimension Patterns
The following figure illustrates the stages of processing in a dimension and the components utilized in developing a dimension load process.
The data load processes comprise of a package and one or more mapping. The package uses B1_JOB_ID mandatory variable as an input. It is used to pass the current job identifier.
The first mapping is usually a view where filters are applied based on the variables to exclude data that does not fall into the specified range. It reduces the data processed in one execution. The data from the view is first inserted into a staging table. The staging table includes the following:
• Source natural key columns.
• Columns mapped to target or used for filters.
• Columns marked for user extension (for dimensions these are UDF codes and description columns).
• Job identifier to segregate the data from multiple parallel executions of a data load process.
The UDX table (refer to the UDX Processing section for more details about UDX tables) is created only if the CM procedure has been configured for the entity. It includes the following:
• Source natural key columns.
• Columns marked for user extension (for dimensions these are UDF codes and description columns).
• Job identifier to segregate the data from multiple parallel executions of a data load process.
The data is finally loaded into the target dimension.