The DrillThrough operations are extremely useful when the user wants to look at the underlying causes for some abnormally high or low outputs in specific fields. For example, let us say a store is doing extremely well and the sales in that store is showing increase in a particular month. The Administrator would like to examine the reasons for this high sales figure and find out if there is any specific reason why this store is doing extremely well when compared to other stores in the region. He will then look at details of the sales in the particular store. Let us say that the sales in this store became abnormally high during the month of July---specifically--the sale of umbrellas. On deeper examination it was found that the sale was at a peak during the last fortnight of July. Then the Administrator could discover that during this period there was an unusual spell of rains and hence the increase in the sale of umbrellas. He would then conclude that the sale was a seasonal one and not to be examined further. However, if the increase in sales happened due to the extraordinary efforts or campaigns undertaken by the store, he would definitely like to study it and take action to replicate the campaign in the other stores and for other months if possible.

Nevertheless, it must be noted that he data in the Data warehouse goes through a process of Extraction, Transformation and loading. Pre-aggregations of data and summarization of data takes place before the data is ready for use in fact tables. The cubes are constructed using the fact tables and dimensions in the database. Therefore, the data that is actually available for Drill through in the Data warehouse is the data that has been pre aggregated, transformed or summarized in different ways. The data that is displayed at drillthrough is actually the data in the fact table. Though this makes for speed and accuracy of presentation, it loses out on the depth of detail that is available in the Online Transaction Processing (OLTP) database. Analysis of the kind illustrated above may not always be possible with the data available in the fact tables.

Drillthrough operations also require that the Client supports the Drillthrough operations. Clients such as the Cube Browser or applications that use COM interface of SQL Server 2000 Analysis Services libraries will be able to Drillthrough to the raw data from the OLAP cube to which they are connected.

The table rows have to be evaluated to the calculated value of the selected cube cell if the drillthrough result set is to be derived. For instance if a selected cube is a three dimensional cube and the intersection of the dimensions puts the values 1998 and Mexico into the cell and is sliced by All Customers and the measure used is Sales, the resulting data set from a drillthrough would be generated only if the fact table rows relate to both 1998 and Mexico--the time dimension table rows store 1998; the Location dimension table rows must store Mexico and all rows of the Customer dimension table. The result set will be summarized by the joins in the Cube’s schema.

Drillthrough is not supported for cells that have values based on expressions such as calculated member cells or ones that are associated with Custom Member formulas. Cells having values for the level members of a mining model dimension also cannot be drilled through.

Permissions to Drillthrough are granted by the Administrator in the Cube Role Manager. To enable or disable Drillthrough of a cube, users will have to Click the Drillthrough options in the Tools menu of the Cube Editor or the Virtual Cube Editor and check or uncheck the Enable Drillthrough check box. The Columns tab will permit the user specify the columns on which the drillthrough will have to be displayed. Drillthrough can be tested by right clicking on a cell in the cube and Clicking on Drill Through.

To specify drillthrough options for a regular cube

In the Analysis Manager tree pane, under the database that contains the cube, expand the Cubes folder.

Right-click the cube, and then click Edit.

On the Tools menu, click Drillthrough Options. (If this option is not available, the cube structure is not valid. The existence of the problem can be ascertained by clicking Validate Cube Structure on the Tools menu.)

In the Drillthrough Options dialog box, perform one of the following actions:

To enable drillthrough, select the Enable drillthrough check box.

To disable drillthrough, clear the Enable drillthrough check box. Go to Step 7.

In the Columns tab, select the columns to be displayed when drillthrough is executed.
Please note that selecting a column to which a cube role is denied access and then granting the cube role the ability to drill through, security exposure becomes an issue. (Access to a column can be denied by using dimension security or cell security). When the end users of the cube role drill through, they can access the denied column.

(Optional.) In the Filter tab, specify a filter (WHERE clause expression) to limit the result set returned by drillthrough. Do not include the keyword WHERE. For example, to limit the result set to fourth-quarter data, specify:

"shipment_facts"."quarter"=4

Click OK.

In Cube Editor, on the File menu, click Save.