In some situations, you may need to create a sequence of datasets, in which the output of one recipe becomes the input of another recipe. 

Potential uses:

  1. You may want to handle data cleanup tasks in one set, before that data is made available to other users for customization for their needs.
  2. Columns or rows of data may need to be dropped before the dataset is made available to other users.
  3. You may want to have different individuals working on each phase of the data transformation process. For example, one individual may be responsible for cleansing the data, while another may be responsible for transforming the data into final format.

Depending on your situation, you can apply one of the following solutions.

Chain Recipes in Same Flow

Within a flow, you can chain together recipes. For example, you may wish to use the first recipe for cleansing and then second recipe for transforming. This method is useful if you are using a single imported dataset for multiple types of transformations within the same flow.


  1. Click the imported dataset. Click Add new recipe.
  2. Click the new recipe. Name it, Cleanse.
  3. With the new recipe selected, click Add new recipe.
  4. Click the new recipe. Name it, Transform.

The output of Cleanse recipe becomes the input of Transform recipe. 

Chained recipes


Create Reference Objects

If you need to make the output of a recipe available in other flows, you can create a reference object. This reference is available in other flows that you control. 


  1. In Flow View, select the recipe whose output you wish to make available to other flows. 
  2. Click the Create Reference icon:

    Create reference object
  3. To use it in one of your other flows, click Add to Flow....
  4. In the target flow, the reference object appears as a reference dataset. It works like an imported dataset with the following considerations.

Key Considerations:

Create Imported Dataset from Output

If any of the above considerations are a concern, you can create an imported dataset from the job results of flow #1. 

In the Job Details page, click the Output Destinations tab. For the generated output, select Create imported dataset from its context menu.

NOTE: When the new dataset is created, it is accessible only to the creator. Datasets can be shared with other collaborators. For more information, see Overview of Sharing.

From the results of wrangling your first dataset, you can create a new dataset. This dataset is wrangled in a separate recipe, the output of which can become a third dataset. In this manner, you can create sequences of datasets. 

Key Considerations:

See Job Details Page.