A transform node in a Redivis project allows users to query one or more tables, using the variables and values to generate a single output table. In this example, we will use the CMS Public Outpatient Charges dataset to walk through the different components of a transform.

All of the following transform steps are optional, except that a transform must have at least one outputted variable to be valid.

Source table

All transform nodes have one source table. You can join multiple tables on in addition, but the source table designation will affect how the joins execute. You can change the source table at the top of a transform. Note that if the new source table has different variables your transform might become invalid until you update it.

1. Build

The first logically executed step of the transform allows you to join tables, create variables, and filter rows. Functionality specified in the Build section will operate on the source table specified at the top of the section, directly upstream of the transform.

2. Partition

The second logically executed step of the transform allows you to create partitioned variables using the output of the Build step. If partitioned variables are created, rows can then be filtered using the values of the partitioned variables.

3. Output

The third logically executed step acts on the outputs of the Build and Partition steps, allowing you to drop duplicates, order rows, and limit output table size.

Run the transform!

To run the transform, click the Run button to the top right of the transform.

This will performs the query constructed in the Variable selector, Build, Partition and Output sections above on the input table (and any other tables joined in), and populates the output table with variables and values according to your specifications.