The aggregate operator takes an input array, list of one or more aggregate function calls, and (optionally) a list of dimensions along which the aggregates are to be computed.  It returns an array containing the computed aggregate values.

Synopsis

aggregate(array, aggregate_1
          [, aggregate_2,... aggregate_N]
          [,dimension_1, dimension_2,...]);

Summary

SciDB aggregates are functions of several values that produce a scalar value. Used together with the aggregate and redimension operators, they compute data aggregations, optionally grouped along a specified dimension or dimensions.
The aggregate operator takes the following arguments: 

The aggregate operator returns a SciDB array containing the aggregation result as a new attribute. The name of this attribute is constructed as follows: inputAttribute_aggregateFunction, where:

For example, assume an array A with an attribute sales. The name for the result attribute of aggregate(A,sum(sales)); is 'sales_sum'.

Example

To create a sparse array and then performs some aggregation on it, do the following:

1. Create an array containing values along the diagonal, and empty cells everywhere else:

   AFL% store(redimension (
                    apply (
                          build(<val:double>[i1=0:3; j1=0:3], i1),
                    i, iif(i1=j1, i1,null), j, iif(i1=j1,i1,null)),
              <val:double>[i=0:3; j=0:3]),A);
   {i,j} val
   {0,0} 0
   {1,1} 1
   {2,2} 2
   {3,3} 3

2. Use the aggregate operator to count the number of non-empty cells:
   

   AFL% aggregate(A, count(val));
   {i} val_count
   {0} 4

    

3. Use the aggregate operator to calculate the sum of all cells in the array:
   

   AFL% aggregate(A, sum(val));
   {i} val_sum
   {0} 6