Updating and Refreshing Cube Data

Updating and Refreshing Cube Data

Many of the changes you make within Analysis Manager and all of the changes to a cube's source data require the cube to be processed in order for the changes to be reflected in the cube's data.

You can process a cube in the following ways:

• Incremental update

  Adds new data to a partition in the cube and updates aggregations. This method does not process changes to a cube's structure (measures, dimensions, and so on) or changes to its existing source data. An incremental update creates a temporary partition from the new data and merges it into an existing partition.

• Refresh data

  Clears and reloads a cube's data and recalculates its aggregations. Use this method if the cube's source data has changed but its structure has not.

• Full process

  Completely restructures a cube based on its current definition and then recalculates its data.

For more information about the kinds of changes processed by the preceding methods, see Processing Cubes.

Because an incremental update creates a temporary partition from the new data and merges it into an existing partition, it is necessary to understand the special considerations that apply to partitions before performing an incremental update. For more information, see Incremental Updates and Partitions.

Referential integrity of the data warehouse is not verified by Microsoft® SQL Server™ 2000 Analysis Services. So, for example, if the cube's (or one of its partition's) fact table contains foreign key values that are not present in a joined dimension table's primary key column, the rows containing those values are not processed. In this case, processing does not produce an error message, but the cube contains incomplete and therefore inaccurate data.

To incrementally update a cube, use the Process a Cube dialog box, the Incremental Update Wizard, and the Process dialog box.

To incrementally update a cube