Working with API Data Transformation in Workflows – Assembled

Overview

This article covers how to retrieve data via GET requests, transform and manipulate arrays (including editing specific items), and submit the updated data through POST requests in Assembled workflows.

Core Concepts

Variables and Data Flow

• Variables are the key to passing data between workflow steps
• Data from GET requests can be stored as variables and referenced throughout your workflow
• Use the {{variable_name}} syntax to reference variables in subsequent steps

The "Analyze case" Step

The "Analyze case" step is your most powerful tool for data manipulation. It can:

• Extract specific values from API responses
• Transform data structures
• Edit array items
• Parse complex JSON objects
• Create new variables from existing data

Step-by-Step Implementation

1. GET Request - Retrieve Data

Set up a Custom API step to fetch your data:

• Configure your GET endpoint
• Map the response to output variables

Note: If you want to capture the full response, use JSON path "." to return the entire payload as a single variable.

2. Analyze Case - Transform and Edit

Use an Analyze case step to manipulate your data

Variable configuration

• Variable Name: Something descriptive and meaningful
• Variable Type: Text
• Description: Clear, specific instructions for the transformation you want to perform

Example variable descriptions

Edit single array item:

Using the data from {{api_response}}, find the artist with id: 2 and update their role_ids to [1, 3, 5]. Keep all other artists unchanged. Return the complete updated artists array.

Transform keys + edit array:

Using {{api_response}}:
1. Rename 'role_ids' to 'assigned_roles' in all artist objects
2. Find the artist named "Artist A" and change their status to "featured"
3. Add a new field 'updated_at' with {{todays_date}} to all artists
4. Return the complete transformed array

Multiple array operations:

Transform {{api_response}} by:
1. Remove any artists where status = "inactive"
2. For all remaining artists, add role_id 99 to their existing role_ids
3. Sort the array by artist name alphabetically
4. Return the final updated artists array

Conditional transformations:

Using {{api_response}}:
- If artist has role_ids [1,2], change to [1,2,3,4]
- If artist name contains "Senior", add role_id 99
- Otherwise, keep existing role_ids unchanged
Return the updated artists array

Data type conversions:

Transform {{api_response}} by:
1. Convert all id fields from strings to numbers
2. Change date fields from "YYYY-MM-DD" to "MM/DD/YYYY" format
3. Ensure role_ids is always an array (convert single values)
4. Return the converted data

Tips for array manipulation

• Be specific about which item to edit (use ID, name, or position)
• Provide an example of the expected output format
• You can edit multiple items in a single step
• Text variables can return complex JSON structures

3. POST Request - Submit Updated Data

Create another Custom API step for your POST request. In the body, reference your transformed variable, for example:

{
 "artists": "{{updated_artists}}"
}

Testing Your Transformations

Use the Preview button in the workflow toolbar and run the workflow with a previous case ID, test email, or test chat. This lets you check how data flows step by step:

1. Verify that {{api_response}} contains the expected data
2. Confirm that {{updated_data}} reflects the correct transformations
3. Validate the POST payload

Common validation checks:

• Array structure is preserved
• Required fields aren't accidentally removed
• Data types match API expectations
• IDs and references remain valid

Best Practices

Writing clear instructions

• Be specific about what to modify
• Use examples of expected output
• Break complex changes into steps
• Call out required data types if the API is strict

Variable naming

• Use descriptive names (updated_artists not data)
• Match your API terminology
• Use multiple variables for more complex transformations

Error handling

• Add conditional branches after API calls
• Validate data exists before transforming
• Include fallbacks for missing/malformed data

Troubleshooting Common Issues

Variable not found

• Ensure variable names match exactly between steps
• Check that the GET request completed successfully

JSON format errors

• Be explicit about expected format in variable descriptions
• Use the preview feature to validate the outputted structure

Array reference issues

• Use unique identifiers (ID, email) rather than array positions
• Handle cases where target items might not exist

Getting Help

For complex transformations or specific use cases not covered here, reach out to your AI Deployment Strategist team for personalized guidance and implementation support.