Working with Aggregation Methods

What are the aggregation methods?

The numerical aggregation methods provided by the OLAP API include the methods in the following table. You can also write your own aggregation functions as described in "Creating Your own Numerical Functions".

Method	Description
average	Calculates the average of the values of a NumberSource.
maximum	Identifies the largest value of a NumberSource.
minimum	Identifies the smallest value of a NumberSource.
total	Calculates the sum of the values of a NumberSource.

There are two different versions of each of the numerical aggregation methods. One version excludes all null values when making its calculations. The other version allows you to specify whether or not you want null values included in the calculation. Each version returns a new NumberSource that, for each set of input values, has an element whose value is the sum of all of the elements in the original NumberSource that have the same set of input values

How do the aggregations methods work?

Standard numerical methods like stdev() work on each element in a NumberSource. An aggregation method is a method like total() that uses the values in a series of Source elements to perform its calculations. When a NumberSource does not have any inputs, this method creates a new NumberSource with a single element whose value is the sum of the values of the elements in the original NumberSource.

When a NumberSource has inputs, each set of input values identifies a subset of elements (tuples) that are arranged by the outputs of the NumberSource (if any). In this case, an aggregation method works on each set of uniquely positioned elements. In other words, when a NumberSource has inputs, an aggregation method calculates the result of the function for each subset.

For more information on how OLAP API methods determine the position of an element and therefore how they determine what elements to use when calculating the values of aggregation methods, see "Finding the position of elements".

Example: Calculating the sum of the elements when a Source does not have inputs

Assume that you have the Source named unitsSoldByCountry (shown below) whose elements are the total number of units for each product sold for each country.

Output	Output	Element
products	countries	Integer
395	Australia	1300
395	United States	800
...	...
49780	Australia	10050
49780	United States	50
49780	...

Now assume that you want to total these values. Since both products and countries are outputs, when you issue the code shown below, the new NumberSource calculates the total number of units sold for all products in all countries.

NumberSource totalUnitsSold = unitsSoldyByCountry.total();

The new NumberSource called totalUnitsSold has only a single element that is the total of the values of the elements of unitsSoldByCountry.

Element
Integer
11350

Example: Calculating the sum of the elements when a Source has inputs

Assume that you have the Source named unitsSoldByCountry (shown below) whose elements are the total number of units for each product sold for each country.

Output	Input	Element
countries	products	Integer
Australia	395	1300
Australia	49780	10050
...	...	...
United States	49780	50
United States	395	800
...	...

Now assume that you total these values. Since product is input, when you issue the code shown below, the new NumberSource calculates the total number of units sold for all products in each countries;. It does not calculate the total for all products in all countries.

NumberSource totalUnitsSoldByCountry = unitsSoldByCountry.total();

The new NumberSource called totalUnitsSoldByCountry has the structure and values shown below.

Input	Elements
countries	Integer
Australia	11350
...	...
United States	850
...	...