Dimensions

A dimension provides contextual information to measures in the BI Gateway model. Each dimension typically stores related information. For example, a product dimension can store the product ID, product name, and product description. Product ID, product name, and product description are the fields of the product dimension. Dimension fields are represented by columns of the dimension table in the BI Gateway datastore. You can map the dimension fields to source data item fields.

In the BI Gateway model, you must provide at least one dimension key for each dimension. The dimension key is a field or a group of fields that uniquely identifies each record of a dimension. You can include a maximum of eight fields in a given dimension key. It is critical to ensure that your dimension key combination will generate unique records. If a duplicable row is found during execution, the complete dimension will be invalidated.

You must specify a refresh rate for a dimension. The BI Gateway services extract data from the data sources, transform it, and populate the respective dimension table. Thereafter, the dimension is periodically refreshed as per the refresh rate. You can build a filter expression to filter the required transformed data from external data sources and populate the dimension table in the datastore.

You can add and configure dimension links in a dimension. Dimension links help you relate dimensions between each other. When defining a measure, you can add linked dimensions to the measure context even if the measure source data has no direct reference to the linked dimension. Measure source data will be inferred from the dimension that defined the dimension link.

You can also configure a dimension as time-slicing. Such dimensions define the time periods for which measure calculations will be computed. For example, a shift schedule or Work Order dimension can be configured as a time-slicing dimension through the fields that identify the start and end times of the period they represent.

The first record in the dimension table represents an unknown record. The unknown record is created automatically. It consists of a primary key field equal to 0 and null values in all the associated source data item fields. The unknown record is used in the following scenarios:

• At run time, the system looks up for each dimension link and measure relationship with the values of the source field(s) used in the link or relationship. These values are then associated with a dimension key value. When the source field value is "Null" or does not exist in the dimension, then the corresponding record or aggregate will reference the unknown dimension record.

• When you relate a constant value to the dimension key ("Null" is also a constant), then the constant value is attached to all of the aggregates from the source system after the calculation of the aggregates. The resulting dimension key in the measure table points to the relevant dimension record. If the constant value does not exist in the dimension or is "Null", then the corresponding record will reference the unknown dimension record.