JSON: Saving Structured Data

Why this matters

Lists and dicts are great while your program runs, but how do you save one to a file or send it over the internet? JSON is the standard text format for exactly that. Web APIs, settings files and many apps speak it, and it maps almost directly onto Python’s dicts, lists, strings, numbers, True/False and None.

The idea

import json

settings = {"name": "Ada", "scores": [90, 85], "admin": True, "nickname": None}

text = json.dumps(settings)               # Python object -> JSON text (a string)
back = json.loads(text)                   # JSON text -> Python object
print(back == settings)                   # True

The s stands for “string”. The versions without s work directly with files:

with open("settings.json", "w") as file:
    json.dump(settings, file, indent=2)   # indent makes it readable

with open("settings.json") as file:
    loaded = json.load(file)

JSON uses true, false and null where Python has True, False and None, and the module converts them for you. Keys must be text (a dict key 1 becomes "1"), and only basic types are allowed: you cannot save a set, a function or your own objects directly.

sort_keys=True makes the output order predictable, which is handy for comparisons.

Try it

json.dumps a dict that contains a list and a None. Then json.loads it back and print its type. What happens when you try to dump a set?

Your turn

Write three functions (they receive file paths, so the checks can use temporary files):

• to_json(data): the JSON text for data, indented by 2 spaces with keys sorted.
• save_json(path, data): write data to the file as JSON (indent 2).
• load_json(path, default): read and return the data from the file, or return default if the file does not exist.

Recap

• json.dumps / json.loads convert to and from text; json.dump / json.load work with files.
• Only basic types survive: dicts (text keys), lists, text, numbers, booleans and None.
• Use indent= for people to read and sort_keys=True for stable output.

Common mistakes

• Mixing up dump and dumps: one takes a file, the other returns a string.
• Expecting tuples or sets back. Tuples become lists; sets cannot be saved.
• Loading a missing or empty file, which is an error: check exists() or handle the problem.