Flattening Complex JSON | Dataford Interview Questions

Write a function to transform a complex JSON structure into a flattened relational format.

Implement flatten_json(data), where data is a JSON object containing nested objects and arrays. Return a list of dictionaries with dotted keys. Arrays of objects must produce one output row per object, repeating parent fields; multiple nested arrays expand as a Cartesian product. Arrays of primitive values remain unchanged in their flattened field.

Example: {"id": 1, "items": [{"sku": "A"}, {"sku": "B"}]} becomes [{"id": 1, "items.sku": "A"}, {"id": 1, "items.sku": "B"}]. Preserve input order and do not mutate the input.

Constraints

data is a JSON object composed only of dictionaries, lists, strings, numbers, booleans, and null
Nested object keys are unique within their object
Arrays of objects use a consistent object shape for grading
Preserve dictionary insertion order and array order
The input may contain multiple nested arrays, which expand as a Cartesian product

Function Signature

def flatten_json(data):

Your question is Flattening Complex JSON. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Write a function to transform a complex JSON structure into a flattened relational format.

Example: {"id": 1, "items": [{"sku": "A"}, {"sku": "B"}]} becomes [{"id": 1, "items.sku": "A"}, {"id": 1, "items.sku": "B"}]. Preserve input order and do not mutate the input.

Constraints

data is a JSON object composed only of dictionaries, lists, strings, numbers, booleans, and null
Nested object keys are unique within their object
Arrays of objects use a consistent object shape for grading
Preserve dictionary insertion order and array order
The input may contain multiple nested arrays, which expand as a Cartesian product

Function Signature

def flatten_json(data):

Your question is Flattening Complex JSON. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.