Python Dict vs JSON: What's Different and Why It Matters

Python dicts and JSON look almost identical at a glance. Both use curly braces, both have key-value pairs, both support nesting. This surface similarity is exactly why the differences cause so many bugs — developers assume they're the same thing until something breaks in production.

They are not the same thing. Here's everything you need to know.

The Syntax Differences

Quoting

Python allows both single and double quotes for strings:

{'name': 'Alice'} # valid Python

{"name": "Alice"} # also valid Python

JSON requires double quotes, always:

{"name": "Alice"} # valid JSON

{'name': 'Alice'} # INVALID JSON

Booleans

Python: True, False (capital T and F)

JSON: true, false (lowercase)

{'active': True, 'verified': False} # Python

{"active": true, "verified": false} # JSON

Null / None

Python: None

JSON: null

{'score': None} # Python

{"score": null} # JSON

Tuples

Python has tuples; JSON has no equivalent:

{'coords': (40.7128, -74.0060)} # Python tuple

{"coords": [40.7128, -74.0060]} # JSON array (closest equivalent)

Key types

Python dict keys can be any hashable type: strings, integers, tuples:

{1: 'one', (0, 0): 'origin'} # valid Python dict

JSON keys must be strings:

{"1": "one"} # only string keys allowed

The Data Model Differences

Python dicts are a runtime data structure — they live in memory as Python objects. JSON is a serialisation format — it's text.

This distinction matters because:

- Python dicts can hold any Python object as a value (datetime, Decimal, custom class instances, lambda functions). JSON can only hold strings, numbers, booleans, null, arrays, and objects.

- Python dicts can have tuple or integer keys. JSON objects require string keys.

- Python dicts are ordered (as of Python 3.7). JSON objects are officially unordered (though most parsers preserve insertion order in practice).

Python's JSON Module

The standard way to convert between the two is Python's `json` module:

import json

data = {'name': 'Alice', 'active': True, 'score': None}

json_string = json.dumps(data)

# '{"name": "Alice", "active": true, "score": null}'

parsed = json.loads(json_string)

# {'name': 'Alice', 'active': True, 'score': None}

The module handles True → true, False → false, None → null automatically in both directions.

What json.dumps Can't Handle

By default, json.dumps raises a TypeError for non-JSON-serialisable values:

import json

from datetime import datetime

data = {'created_at': datetime.now()}

json.dumps(data)

# TypeError: Object of type datetime is not JSON serializable

Solutions:

- Use a custom encoder: json.dumps(data, default=str)

- Use a library like pydantic or dataclasses-json that handles serialisation

- Use orjson or ujson for faster serialisation with more type support

When to Use Python Dict vs JSON

Use Python dict when:

- Working within a Python codebase — pass dicts between functions, store in memory, manipulate in place

- You need Python-specific types (datetime, Decimal, Enum, set)

- Performance matters — dict operations are faster than parsing/serialising JSON

Use JSON when:

- Sending data over a network (HTTP APIs, WebSockets)

- Writing to disk for other systems to read

- Communicating between services in different languages

- Storing structured data in a database as a text or jsonb field

Django and Flask: The Serialisation Gap

The most common Python dict → JSON friction point is in web frameworks. Django's ORM returns QuerySet objects; Flask/FastAPI route handlers receive request.json as dicts.

The gap manifests at the boundary — turning database objects into API responses. This is why Django REST Framework serializers, Flask-Marshmallow, and FastAPI's Pydantic models exist: they bridge the Python object model to JSON's stricter type system.

When debugging these boundaries — inspecting what a model instance contains, or what a request payload looks like — you're constantly copying Python dict repr output and wanting JSON. That's what DevConvert's Python Dict converter handles.