Example 3: Selection views based on conditions

Using condition functions (predicates), we can filter away uninteresting data and keep only interesting data. In physics codes, this process is often called cutting on a predicate.


Conditions on 1-d matrices (vectors)

// the naming shortcut (alias) saves some keystrokes:
cern.jet.math.Functions F = cern.jet.math.Functions.functions;

double[] v1 = {0, 1, 2, 3};
DoubleMatrix1D matrix = new DenseDoubleMatrix1D(v1);
// 0 1 2 3 

// view all cells for which holds: lower <= value <= upper
final double lower = 0.2;
final double upper = 2.5
matrix.viewSelection(F.isBetween(lower, upper)); 
// --> 1 2

// equivalent, but less concise:
matrix.viewSelection( 
	new DoubleProcedure() {
		public final boolean apply(double a) { return lower <= a && a <= upper; }
	}
);
// --> 1 2

// view all cells with even value
matrix.viewSelection( 
	new DoubleProcedure() {
		public final boolean apply(double a) { return a % 2 == 0; }
	}
);
// --> 0 2

// sum of all cells for which holds: lower <= value <= upper
double sum = matrix.viewSelection(F.isBetween(lower, upper)).zSum(); 
// --> 3

// equivalent: double sum = matrix.viewSelection(F.isBetween(lower, upper)).aggregate(F.plus,F.identity);

Conditions on 2-d matrices

// view all rows which have a value < threshold in the first column (representing "age")
final double threshold = 16;
matrix.viewSelection( 
	new DoubleMatrix1DProcedure() {
		public final boolean apply(DoubleMatrix1D m) { return m.get(0) < threshold; }
	}
);

// view all rows with RMS < threshold.
// the RMS (Root-Mean-Square) is a measure of the average "size" of the elements of a data sequence. final double threshold = 0.5; matrix.viewSelection( new DoubleMatrix1DProcedure() { public final boolean apply(DoubleMatrix1D m) { return Math.sqrt(m.aggregate(F.plus,F.square) / m.size()) < threshold; } } );

Conditions on 3-d matrices

// view all slices which have an aggregate sum > 1000
matrix.viewSelection( 
	new DoubleMatrix2DProcedure() {
		public final boolean apply(DoubleMatrix2D m) { return m.zSum() > 1000; }
	}
);