The Data Refresh Methods page identifies the objects included in the build, the preferred build method, and the cube script that defines the steps of the build.
The table contains a row for each object included in the build. It has the following columns:
Order: The sequence in which the objects are processed. Unless there are dependencies among the cubes that require a particular processing order, select the Automatic Order check box. Deselect Automatic Order if you want to change the processing order of the cubes using the up- and down-arrows. This column appears only when the Automatic Order box is not selected.
Object Name: The name of a dimension or a cube included in the build.
Object Type: Dimension or Cube.
Synchronize: Updates the dimension so that its members are the same as those in the source table, by adding new members and deleting others. Otherwise, new members are added but none are deleted. You can change this setting only when the dimension is not enabled as a materialized view.
Refresh Method: The method used to refresh and aggregate the data. Cubes that have been enabled as cube materialized views use different methods than other cubes. The list contains only those methods applicable to a particular cube. Analytic Workspace Manager uses the selected method, but will revert to a different method if necessary to complete the build. Choose from among these methods:
Complete: Clears all data from the cube, then loads and aggregates all the data from the source tables. All cubes can use this method.
Fast: Uses the materialized view log tables to identify, load, and aggregate only the new and changed data from the source tables. Cubes defined as a fast refresh or a rewrite materialized view can use this method.
Force: Uses the Fast method if possible; otherwise, the Complete method.
Partition Change Tracking: Clears, loads, and aggregates only the values from an altered partition in the source tables.
Fast Solve: Loads all the detail data from the source tables, then aggregates only the new values. Compressed cubes and cube materialized views can use this method.
Cube Script: Lists the available cube scripts, from which you can select one to use for this build. The selected cube script can affect the available methods. Cube materialized views always use the SYS_DEFAULT script.
Automatic Order
Controls whether you or the OLAP engine defines the processing order. Deselect this check box when the objects have dependencies and must be processed in a particular order. Then use the up- and down-arrows to list the objects in the correct order. Otherwise, select this option to enable parallel processing and to allow the OLAP engine to select the most efficient processing order. The up- and down-arrows will be disabled and the Order column will be hidden.
Copyright © 2003, 2007 Oracle. All rights reserved.