Lecture 15

JSON

MCS 260 Fall 2021
David Dumas

Reminders

Hints

• In the terminal:
  • Tab = autocomplete
  • ↑ = previous command
  • ls lists files in current directory
• You can download sample scripts from Github using the "Raw" button, then File>Save.

JSON

JSON stands for JavaScript object notation. It is a format for storing various types of data in text files. Many languages can read and write this format.

Python has a module for reading and writing JSON.

By using that module, you can move various types of data in and out of files cleanly, without needing to handle the read/write details yourself.

JSON types

Supported basic types:

• string — must use double quotes.
• number — float, int, other? Up to reader.
• boolean — lower case names true, false.
• null — similar to Python's None.

JSON types

Supported composite types:

• array — ordered sequence of values like Python list. Surrounded by square brackets, values separated by commas.
• object — associative array like Python dict. Surrounded by curly braces, comma separator. Keys must be strings.

A JSON file must contain a single value. Most often it is an object or array.

Writing JSON

Import the module json to make JSON function available.

You'll need to open a file yourself; JSON functions expect a file object.

Use json.dump(val,f) to write val to file object f as JSON.

Python to JSON type conversion table:

• dict → object
• list or tuple → array
• int or float → number
• bool → boolean
• None → null

JSON string

If you want to handle writing yourself, you can ask the JSON module to convert a value to a JSON string, e.g.

json.dumps(val) # make JSON string out of val

Reading JSON

Use json.load(f) to interpret contents of file object f as JSON and return the decoded result.

json.loads(text) will instead process string text as JSON.

Not supported in JSON

• Complex numbers
• Date/time types
• Distinctions between:
  • int and float^*
  • tuple and list
• Comments

* But Python's json module will try to guess when reading.

Example

Let's modify wordstats3file.py so that it writes a dictionary of statistics to a JSON file.

That way, another program can read and use the report easily without concern for formatting details.

Example

One of the items in that list is a data file listing episodes of the Netflix TV show Stranger Things. The link to it is actually broken, but the list is available at:

https://api.tvmaze.com/shows/2993/episodes

Let's download it and poke around using Python.