Advanced Metric Options dialog box: VLDB Properties

VLDB (Very Large DataBase) properties allow you to alter the syntax of a SQL statement and take advantage of the unique, database-specific optimizations. The following metric-specific VLDB properties can be set at the metric level:

Null checking for Analytical Engine: Determines whether a null value is interpreted as zero when the Analytical Engine performs calculations.

Subtotal Dimensionality Aware: Enables subtotaling based on the dimensionality of a metric.

Metric Validation: Determines whether dynamic sourcing is available for metrics. In general, if metrics use outer joins, accurate data can be returned to reports from Intelligent Cubes through dynamic sourcing. However, if metrics use inner joins, which is a more common join type, you should verify that the metric data can be correctly represented through dynamic sourcing. You should disable dynamic sourcing if your metric data is not modeled to support dynamic sourcing. The inclusion of null values in fact tables that contain your metric data can cause incorrect data to be returned to reports from Intelligent Cubes through dynamic sourcing.

Integer Constants in Metric: Determines whether to add a .0 after an integer.

Metric Join Type: Defines the default action for joining the metric to other metrics. The options are:

- Inner join, which includes only rows that have data present for all metrics across that row in a report
- Outer join, which shows all rows for the metric, regardless of other metrics on the report

Null Check: Determines how to handle arithmetic operations with null values. The Analytical Engine can either check for null values in all queries or in temporary table joins only.

Transformable Agg Metric: Defines what metrics should be used to perform transformations on compound metrics that use nested aggregation.

Zero Check: Determines how to handle division by zero or when to check for zeroes in the denominator during division calculations. When a zero check is performed, a zero in the denominator of a division calculation is changed to a null value. The options are:

- Do nothing, which means that the database rather than the Analytical Engine handles division by a zero
- Check for zero in all queries
- Check for zero in temp table join only

Count Distinct: Determines how to handle queries performed on multiple partitioned tables which return a distinct count of values. A distinct count of values allows you to return information such as how many distinct types of items were sold on a given day. Selecting the correct option for your project setup can help improve performance.

The default setting for each VLDB property is inherited. You can set additional VLDB properties at other levels, such as the report and project levels.

To access the Advanced Metric Options: VLDB Properties dialog box:

Prerequisites

This procedure assumes that you have created a metric. For steps, see Creating a metric.

If the metric is not already open, click the metric to edit it. Either the Function Editor or Formula Editor opens, depending on the type of metric.
Click the Options button. This displays the Advanced Metric Options dialog box.
In the list of categories on the left, click VLDB Properties.