Specifying Dimension Order for Financial Data Items

Importance of dimension order for financial data items

When you create a stored financial data item, you must pay attention to the order in which you select the dimensions that comprise the financial data item. The dimension order determines how data is stored in Financial Analyzer, which in turn affects performance and storage in the Financial Analyzer database.

Recommendations for dimension order for financial data items

The order that you specify for the dimensions in your financial data items should depend on whether the data in the financial data item is dense or sparse.

Dimension order recommendation for dense data

When data is dense, the Time dimension should be the last dimension in the financial data item definition.

Dimension order recommendations for sparse data

When data is sparse, Oracle Corporation recommends that you follow these guidelines:

• You must have at least one dense dimension.

• Place dense dimensions first in the financial data item's definition.

• The Time dimension should always be a dense dimension.

• If you have more than one dense dimension (one of them being Time), the Time dimension can either be the last of the dense dimensions or the last of all dimensions.
  • If you will be consistently processing one period's data at a time for a financial data item, placing the Time dimension last may improve load and solve performance.

  • If you will be consistently processing a time series, placing the Time dimension last of the dense dimensions may improve load and solve performance.

• Dimensions identified as sparse should be ordered according to their number of dimension values, with dimensions with more values coming before those with less. For example, if a financial data item has four dimensions: Account Organization, Product, and Time, and the relationship between Account, Organization, and Product is sparse, then the order of the dimensions should be:

  <Time <Account Organization Product>>
  

  where Account, Organization, and Product are in the composite.

  If a financial data item has four dimensions: Account, Organization, Product, and Time, and only the relationship between Organization and Product is sparse, then the order of the dimensions could be:

  <Account Time <Organization Product>>
  

  where Organization and Product are in the composite. Also, one period's data is typically loaded and solved at once.

  Alternatively, the order of the dimensions could be

  <Account <Organization Product> Time>
  

  where Organization and Product are in the composite. Also, one period's data is typically loaded and solved at once.