Why my json result contains two sections with the same name?

Question

I use the following code to get json from the string value, update it and return back as string (simplejson is used):

if text:
    qj = json.loads(text)
else:
    qj = json.loads('{}')
if not qj.get(counter):
    qj[counter] = {}
if not qj.get(counter).get('values'):
    qj[counter]['values'] = []
if not value in qj[counter]['values']: # avoid duplicates
    qj[counter]['values'].append(value)
text = json.dumps(qj)

Why do I get the following result if I run the function with the same counter value several times (4 in the example below):

{"4": {"values": ["test1"]}, "4": {"values": ["test2"]}}

The number of created 4 sections is always 2. If I run the function 3rd time, then first value is updated (not added as I expect).

The expect output should be:

{"4": {"values": ["test1", "test2", "test3"]}}

What is wrong in my code?

Answer 1

If in your first pass you have following json:

{4: {"values": ["test1"]}}

You will have this string after text = json.dumps(qj):

'{"4": {"values": ["test1"]}}'

Now, when loading the json with qj = json.loads(text), you receive:

{'4': {'values': ['test1']}} # 4 has turned to '4'!

Your check if 4 already exists, fails because it is a number.

So to solve your problem, you can simply turn the counter into a string during the test:

if not qj.get(str(counter)):
    ...

The reason for this behaviour is that json does not allow integers as keys, see the definition of a pair in the JSON documentation:

pair

string : value