Overview of the D-Cube

In Analyst, data is stored in a D-Cube. The D-Cube is a multi-dimensional data structure similar to an OLAP or Excel pivot table. Each of the dimensions contains a list of related data. For example, in a four dimensional cube, the first dimension may contain the items in  P&L statement, the second may list the departments, the third show the timescale, and the fourth may show the version (Budget or Actual). In this example, you have the  P&L statement by month, by department, and by version. The D-Cube must have at least two dimensions. If more than two dimensions exist, the other dimensions appear as pages.

The following screenshot depicts a four dimensional D-Cube. The first two dimensions form the rows and columns. The third and fourth dimensions are displayed as pages, in the upper-left corner of the D-Cube. You can switch to different pages in the cube by clicking the page drop-down list and selecting another page.

Creating a D-Cube

Creating a D-Cube takes a few simple steps but it requires a great deal of forethought. Before creating the D-Cube, you must think through the dimensions that will form the cube. You must know what data will populate the D-Cube. You must know what category of D-Lists your selection falls under so that you can prioritize the calculations appropriately. You must know what function the D-Cube performs. Is it an input cube, a calculation cube, a summary or reporting cube, or a staging cube? You must also know in which library the D-Cube will be stored, and whether the D-Cube will be shared or will be exclusive to a model. Finally, you must consider size and performance, especially if you plan to deploy the D-Cube in Contributor.

To create a D-Cube:

1. Click on File | New | D-Cube.
2. In the Create new D-Cube box, select the D-List that makes up the D-Cube.
3. Drop the D-List onto the lower pane.
4. Order the D-List. Click on OK.
5. Enter the D-Cube name, and then click on OK.

By default, the D-Cube opens with the longest D-List as the rows and the timescale as the columns. You can transpose the cube or swap the dimensions by dragging and dropping them. When you close and reopen the D-Cube, it will revert to the default view.

Order of dimensions

It is important to have the proper order of dimensions in the D-Cube. The order of dimensions is important because it determines which calculation takes precedence when calculations from two separate D-Lists intersect. The order also determines the priority of formats when the cell has opposing formats. As a rule, the dimensions must follow this order:

1. Calculation: Calculation D-Lists contain mathematical operations over and above simple addition. Any use of operators, such as multiplication, subtraction, division, or the use of a BiF, make the D-List a calculation D-List.
2. Aggregation: Aggregation D-Lists usually contain a hierarchy of items, such as an organizational hierarchy, a list of products, customers, or cost centers, or a list of items with a simple subtotal.
3. Non-Calculating: Non-Calculating D-Lists contain no calculations.
4. Timescale: Timescale D-Lists contains items that span a period of time, such as months, weeks, or days.
5. Versions: Version D-Lists contain iterations of the data, such as Actual, Budget, Forecast, and Revised Budget.

It is critical to set the proper order of dimensions before building the Contributor application. If you reorder the dimensions after the application is created, all of the data in the D-Cube will be deleted. In addition, reordering dimensions changes the structure of the import tables and, therefore, the way that you import data into Contributor.

Size considerations

The size of the D-Cube is measured by the number of cells in it. The number of cells you can have in a D-Cube is limited only by your computer’s memory. When you build a D-Cube, you should keep in mind the amount of memory available on the computer on which the model will be used. You will not be able to open a large cube if your PC does not have enough memory to handle the data. If the model is deployed in Contributor, size is an even more important consideration. Even with a powerful computer, you will experience performance lag because a large model must pass through the organization’s network. In a wide area network, the problem associated with a large model becomes more apparent.

To determine the number of cells, find the product of the number of items in all of the D-lists in the D-Cube. For instance, if a D-Cube has five D-Lists, which have the following number of items: 8, 1, 10, 5, and 12, the number of cells in the D-Cube will be 4,800. This is 8*1*10*5*12.

You can find the number of cells in the dimension selection box, as illustrated in the following screenshot:

The number of items in a dimension has an inverse impact on the magnitude of size increases for additions to the cube. The fewer the items, the bigger the impact that an addition will have to the size of the cube. In the example of the cube above, if you add one item to a dimension that has only one item, you will double the size of the cube. However, if you add one item to a dimension that has 10 items, you will increase the size of the cube by only ten percent.

One clue that your D-Cube may have a problem with size is the number of dimensions. While there is no limitation on the number of dimensions that a cube can have, having too many of them can become a problem. With a D-Cube that has too many dimensions, any additional item in one of the D-Lists increases the size in an order of magnitude depending on the number of items in the D-List. If possible, keep the number of dimensions to not more than five. If you need to have a D-Cube that provides views of data beyond five dimensions, consider splitting the D-Cube. You are more likely to require less memory if you have several smaller cubes than with one super cube.

With Analyst, you can open a D-Cube that has several million cells on a PC that has 512MB of RAM. However, if you plan to deploy the model in Contributor, the threshold for the number of cells is much lower. An acceptable benchmark is roughly 500,000 cells per e.List. This benchmark is subject to many other factors, such as network latency, client PC memory, and CPU, all of which can contribute to model performance.

Working with the D-Cube

When working with the D-Cube, you have several functionalities that help you to hone in on the information that you want. By default, the D-Cube opens in full view, but you can open only a slice of the cube and save the slice for later viewing.

Opening a full view of the D-Cube

The most straightforward way to view the D-Cube is to open all dimensions. To open a full D-Cube:

1. Click on File | Open | D-Cube.
2. Select the D-Cube. Click on OK.
3. Select the Full option. Click on OK.

Opening a selection of the D-Cube

A selection is a subset of a D-Cube. You can only open the specific items of the dimensions that you want to view. Because the data in a selection is fewer than the full view, less memory is required and the D-Cube opens and recalculates faster. To open a selection of the D-Cube:

1. Click on File | Open | D-Cube.
2. Select the D-Cube. Click on OK.
3. Select the Edit Selection option. This option opens the dimension selection box, where you can select the items that you want to view in the D-Cube slice.

   As shown in the following illustration, the item selection box lets you select the items from a dimension that you want to include in your view. In the item selection box, each tab represents a D-List. To select the items, move the items from the Items available pane (on the left) to the Items included pane (on the right).

   

4. Click on OK.

Viewing different slices of the D-Cube

You can view the D-Cube in a variety of ways. You can open a selection of the D-Cube, or several selections at the same time. If you have a selection open, you can change it to another selection. The ability to switch views makes it easier to work with the D-Cube. For instance, suppose you have a D-Cube that has two versions: Budget and Actual. If you want to enter only budget data, you may want to open only the Budget dimension and keep the Actual dimension closed. You can have multiple selections of the D-Cube open at the same time in the same way that you open the D-Cube. This allows you to arrange the selections next to each other on the same screen.

To modify the views of the D-Cube:

1. With the D-Cube open, click on D-CubeSelections:
   1. New Slice: This option allows you to create a new view of the same D-Cube, with the same selection.
   2. Reselect: Modify the current selection. Selecting this option allows you to add or remove dimensions in the current view.
   3. Save: Save the current selection as a saved selection. This option allows you to save the current view so that you can open it again later.
2. If you select Reselect, move the items that you want to add to or remove from the current view from the Items included to the Items available pane, or vice versa, respectively. Click on OK.

Saving a selection

You can select a view of a D-Cube and then save this selection so that you do not have to reselect the same view the next time that you open the cube. When you save a selection, you create a “saved selection”—an object that is derived from the cube.

To save a selection:

1. Click File | Open | D-Cube.
2. Select the D-Cube.
3. Under Mode, select Edit Selection.
4. Select the D-List items that you want to view in the D-Cube, and move them from the Items available pane to the Items included pane. If you do not select anything in the Items available pane, then all of the items will be included in the selection, including any future additions to the D-List.
5. Under the Save Selections option on the lower-right corner of the dialog box, click on the Save button.
6. Enter the name of the saved selection. Click on OK.

If you have a saved selection, you can load it into the item selection box. The dimension item selection box is used in many functions, including export, D-List imports, and D-Cube allocations (to be discussed later).

Opening the saved selection

Once saved, the selection becomes a separate object that can be copied and shared. However, because it is a subset of the cube, you have to go through the cube to access the saved selection.

To open the saved selection:

1. Click on File | Open | D-Cube.
2. Select the D-Cube.
3. Under Mode, select Saved Selection.
4. Select the Saved Selection. Click on OK