Python 2.7> 3.5 JSON转储“不是字符串”

Question

I am converting my code from Python 2.7 to 3.5 but run into some issues while dumping JSON data into a file.

Originally the code for 2.7 was like this:

backlog_json = json.dumps(backlog, indent=2, sort_keys=True)
f = open(backlog_path,"wt")
f.write(backlog_json)
f.close()

Currently the same code gives the following error:

TypeError: key b'00c4349c-9617-42c3-a0a3-8e16262bcd76' is not a string

It seems to do something with bytes, while I just want all items in my JSON to be strings.

for some reason b' is added to the dict data:

b'f91de40f-cf15-4da6-bee9-0101920c8874':

Answer 1

You need to convert any keys that are byte literals with strings. you can make the swap-with str(bytes_string, 'utf-8') but first you have to find them. If it is obvious where these byte literals exist you can write code to switch them out or you could convert like this...

def reqked(ob):
    if(type(ob) == dict):
        for(k in ob.keys()):
            if(type(k) == bytes):
                ob[str(k, 'utf-8')] = reqked(ob[k])
                del ob[k]
            else:
                ob[k] = reqked(ob[k])
    elif(type(ob) == list):
        for k in range(0,len(ob)):
            ob[k]=reqked(ob[k])
    elif(type(ob) == bytes):
        ob = str(bytes_string, 'utf-8')
    return(ob)

Answer 2

You have to convert your byte data into string one

file = open(file_name, encoding='utf8')
json_data = json.load(file)
file.close()