What is a pipeline?

Pipelines connect Data Sources to Operators and Operators to Actions. Operators are used to process data and run ML algorithms, actions are used to take action based on the results of the operators.

Note that if you try to connect an Action to an Integration you will get a connection error.

Explanation with Example

How does it work?

Each operator and action has input(s) and output(s). While some of the operators/actions and actions allows us to connect a single input, some of them allow us to connect multiple outputs. When an operator is executed, it always creates an output table(s) and then this table(s) can be used in the next operator or action.

In the next two examples, we will explain how operators input and output tables are used.

Explanation with Example #1

Step #1 → We will connect a client data source to a custom SQL operator.

Step#2 → When we click on edit, we can see the editor to write SQL.

Step#3→ On the right panel of the SQL editor, we can verify the client data source is connected to that current operator.

Step#4→ On the next step, let's connect the customer's invoice data source.

Step#5 → When we open the Custom SQL operator's editor by using the edit button, we can verify the 2nd data source is also connected.

The steps explained above can also be seen on the video below.