Practical Real-time Data Processing and Analytics

Functions

The following are characteristics of functions:

• Class has to extend BaseFunction
• This is a partition of the local operation that means no network transfer is involved and is applied to each batch partition independently
• It takes a set of inputs and emits zero or more output
• In output, it emits an output tuple including the original input tuple

Here is the example:

class PerformDiffFunction extends BaseFunction {  @Override  public void execute(TridentTuple tuple, TridentCollector collector) {    int number1 = tuple.getInteger(0);    int number2 = tuple.getInteger(1); if(number2>number1){      collector.emit(new Values(number2-number1));    }  }}

Input:

[1,2][3,4][7,3]

Output:

[1,2,1][3,4,1]