写入 JSON 时如何不转义反斜杠

Question

I am trying to read a JSON file and add a new key,value pair to it. Then write the JSON file back. Example

# Read the JSON and add a new K,V
input_file_path = os.path.join(os.path.dirname(__file__), 'test.json')
input_file = open(input_file_path, 'r')
data = json.load(input_file)
data['key'] = os.environ.get('KEY', '') #<-- Adding a new K,V
input_file.close()

# Write the JSON to tmp.
output_file = open('/tmp/' + 'test.json', 'w')
json.dump(data, output_file, indent=4)
output_file.close()

My input JSON looks like this

{
  "type": "account"
}

My env variable called KEY looks like this -----BEGINKEY-----\\nVgIBMIIE

The final JSON file written to tmp looks like this

{
  "private_key": "-----BEGINKEY-----\\nVgIBMIIE",
  "type": "account"
}

I am not able to figure out why the backslash is escaped? How can I avoid this?

Answer 1

The program is treating your input string as a raw string and so adds the extra \\ . The original '\\' is not actually escaping anything, so for it to be represented in Python as a string, you need to escape it. As you saw this can be problematic however. You can force the string back to unicode format using:

import codecs

raw_val = os.environ.get('KEY', '')
val = codecs.decode(raw_val, 'unicode_escape')
data['key'] = val

Answer 2

If you want to represent -----BEGINKEY-----\\nVgIBMIIE in JSON, it has to go in a double quoted string.

In double quoted strings, backslashes have special meaning. It's not possible to write a backslash character simply by typing "\\" . This would generate an error.

Therefore, the backslash must be escaped. Hence -----BEGINKEY-----\\nVgIBMIIE becomes "-----BEGINKEY-----\\\\nVgIBMIIE" in a JSON string.

If you want to get rid of the backslash, you need to represent

-----BEGINKEY-----
VgIBMIIE

rather than

-----BEGINKEY-----\nVgIBMIIE

These two things are not the same thing.