Use Build Features to create features and establish relationships between data in separate tables. The tool uses primitives to build features from the data you provide.
This tool is part of Alteryx Intelligence Suite. Intelligence Suite requires a separate license and add-on installer to Designer. After you install Designer, install Intelligence Suite and start your free trial.
The Build Features tool has 2 anchors.
- Input anchor: The input anchor connects to the data streams you want to build features from. The 2 angle brackets on the input anchor indicate that it accepts multiple inputs.
- Output anchor: Use the output anchor to pass the data that includes the features you build downstream.
Configure the Tool
To use the Build Features tool, you have to configure options that manage relationships between your data and manage primitives that build the features from your data.
1. Manage Relationships
- Select the Manage Relationships tab. You should see this section by default when you open the Configuration window for the first time.
- Select the Target Table.
- Choose a Primary Key from the dropdown. You only have to choose a primary key for the target table, but selecting a primary key for other tables might help you create relationships, depending on how many relationships you want to create.
- Define parent-child relationships between tables in your data. Choose the Parent table and its Key, as well as the Child table and its Key.
- Select New relationship if you want to add more than one parent-child relationship.
2. Manage Primitives
- After you've created all the relationships you want, go to the Manage Primitives tab.
- Search for the primitives you want to build from the data. To see a list of primitives with their explanations, visit this page.
- Check the box next to those primitives.
- Choose the Table Depth, which specifies how many tables the tool should look at when using aggregation primitives. Those kinds of primitives build features by combining, or aggregating, data from multiple tables.
At a high level, primitives are functions applied to raw data that help build features from it. Those functions can either aggregate or transform the data to build features. Primitives only constrain the input and output of data, so you can apply the same features in many different scenarios. For example, 1 primitive measures the average time between 2 dates. You can apply that primitive in many different scenarios, like to measure the duration of semesters, seasons, or tenures. In that way, a single primitive can be used in different contexts to answer different questions about your data. For more information about how primitives work, visit this page.