json.dump() changes the type from `dict` to `NoneType`

Question

I have a small python program that I use to manipulate JSON files. I find that when I dump the manipulated data back to the JSON file using json.dump() it changes to <type 'NoneType'>. The original type of the data below (type of json_data) is <type 'dict'>. I am storing these JSON documents in elasticsearch and visualising it using Kibana4.Kibana4 is treating newly added integer fields as string. Has someone come across this issue before.

import json                                                                                                                                                                                    

fname = "json_data.txt"
with open(fname, 'r+') as f:
    json_data = json.load(f)
    print(type(json_data))
    #Code to add fields to json files.   
    f.seek(0)
    x = json.dump(json_data,f,ensure_ascii=True)
    print(type(x))

Answer 1

json.dump() has no return value. It writes to the file, not return the dumped object.

As such the default return value for a callable, None is returned instead.

In other words, your code is doing exactly what you asked it to do: read JSON data and parse it, the result is stored in json_data. You then write that Python object back into the file. The return value of json.dump() is irrelevant here, you still have a reference to json_data.

If you wanted a string value containing the JSON object, use json.dumps() (note the s); this returns the resulting JSON string without writing to a file:

fname = "json_data.txt"
with open(fname, 'r') as f:
    json_data = json.load(f)
    print(type(json_data))

json_string = json.dumps(json_data)
print(type(json_string))