Introduction
Analytics Builder provides an environment for connecting blocks together to form models that can process and react to inputs. Analytics Builder uses a few types of values internally, and it is important to understand the differences between these. The following topics cover the distinctions between value types representing continuous-time and discrete-time values, and the pulse type. They also cover some of the details of block implementations around windowing and when blocks generate output.
Summary
In Analytics Builder, the value types float, boolean, and string are used to represent continuous-time values. They have the following properties, which you should consider when writing models or creating custom blocks:
A value wire holds its value until there is new input.
Repeated inputs of the same value should not affect the output (for most cases).
There is no guarantee that a new input will occur within a defined time period.
Contrast these to the pulse type, which represents discrete events and has the following properties:
A pulse represents a single point in time.
Multiple inputs of a pulse have significance, even if there is no difference to any value associated with the pulse.
If a block has multiple input ports for pulses, inputs occurring at different times to different ports will only be
seen one at a time. An input port for a pulse on a block acts as if it is automatically
reset after evaluation.
These properties and the rationale behind them are explored and explained in the following topics. These topics also explain how to handle cases that do not fit into these distinctions, such as discrete numeric measurements.