如何在不引号的情况下转储类似json的字符串？

Question

Instead of this (from json.dumps ):

[
  {
    "created": 581937573,
    "text": "asdf"
  },
  {
    "created": 581937699,
    "text": "asdf"
  }
]

I would like to get this [non-JSON] output:

[
  {
    created: 581937573,
    text: "asdf"
  },
  {
    created: 581937699,
    text: "asdf"
  }
]

If json.dumps had an option to change the quote character for keys, I could just set it to a placeholder and strip them out later. Unfortunately it doesn't appear to have this option... any other clever ideas?

(load from this format is not necessary, but would be interesting if a solution for it exists as well)

Answer 1

re.sub offers a quick fix under certain assumptions:

import re

data = [...]
json_data = json.dumps(data)

with open('file.txt', 'w') as f:
    f.write(re.sub(r'"(.*?)"(?=:)', r'\1', json_data))

file.txt

[
  {
    created: 581937573,
    text: "asdf"
  },
  {
    created: 581937699,
    text: "asdf"
  }
]

These assumptions are

1. your data is small enough that re.sub runs in a reasonable amount of time
2. your data does not contain string values that themselves contain stuff that could match the pattern I've used here.

The pattern effectively looks for all dictionary keys and strips out the quotes.