Was this helpful?
Operators
Logical operators, or more simply operators, are the fundamental building blocks of DataFlow applications. An operator represents some form of data processing. They are connected together to build graphs that describe complex processing.
DataFlow provides a library of operators covering commonly used processes. DataFlow also permits extension by allowing users to create their own operators. An overview of how to create new operators is in Writing an Operator.
Operators can be viewed as having three main sections:
The interface of an operator is defined by the logical ports it declares. These represent the inputs and outputs of the operator. An operator may have any number of inputs and outputs, including none.
The configuration of an operator is defined by the properties it exposes. Most operators describe a general operation which is refined further by property settings. For example, an operator that performs sorting would expose a property to indicate the sort keys to use.
The implementation of an operator describes how to perform the processing. DataFlow allows operators to be defined as graphs, permitting reuse of common graph patterns. To avoid infinite recursion, however, some operators must necessarily provide a more concrete physical implementation.
Of these three pieces, users of an operator need only to be aware of the first two. Creators of new operators must be concerned with all three.
Last modified date: 06/14/2024