Python CSV to JSON W/ Array Output

Pandas solves this really simply. First to read the file

import pandas

df = pandas.read_csv('music.csv', names=("id","Artist","Song", "Album"))

Now you have some options. The quickest way to get a proper json file out of this is simply

df.to_json('file.json', orient='records')

Output:

[{"id":1,"Artist":"Empire of the Sun","Song":"We Are The People","Album":"Walking on a Dream"},{"id":2,"Artist":"M83","Song":"Steve McQueen","Album":"Hurry Up We're Dreaming"}]

This doesn't handle the requirement that you want it all in a "Music" object or the order of the fields, but it does have the benefit of brevity.

To wrap the output in a Music object, we can use to_dict:

import json
with open('file.json', 'w') as f:
    json.dump({'Music': df.to_dict(orient='records')}, f, indent=4)

Output:

{
    "Music": [
        {
            "id": 1,
            "Album": "Walking on a Dream",
            "Artist": "Empire of the Sun",
            "Song": "We Are The People"
        },
        {
            "id": 2,
            "Album": "Hurry Up We're Dreaming",
            "Artist": "M83",
            "Song": "Steve McQueen"
        }
    ]
}

I would advise you to reconsider insisting on a particular order for the fields since the JSON specification clearly states "An object is an unordered set of name/value pairs" (emphasis mine).

Your logic is in the wrong order. json is designed to convert a single object into JSON, recursively. So you should always be thinking in terms of building up a single object before calling dump or dumps.

First collect it into an array:

music = [r for r in reader]

Then put it in a dict:

result = {'Music': music}

Then dump to JSON:

json.dump(result, jsonfile)

Or all in one line:

json.dump({'Music': [r for r in reader]}, jsonfile)

"Ordered" JSON

If you really care about the order of object properties in the JSON (even though you shouldn't), you shouldn't use the DictReader. Instead, use the regular reader and create OrderedDicts yourself:

from collections import OrderedDict

...

reader = csv.Reader(csvfile)
music = [OrderedDict(zip(fieldnames, r)) for r in reader]

Or in a single line again:

json.dump({'Music': [OrderedDict(zip(fieldnames, r)) for r in reader]}, jsonfile)

Other

Also, use context managers for your files to ensure they're closed properly:

with open('music.csv', 'r') as csvfile, open('file.json', 'w') as jsonfile:
    # Rest of your code inside this block

It didn't write to the JSON file in the order I would have liked

The csv.DictReader classes return Python dict objects. Python dictionaries are unordered collections. You have no control over their presentation order.

Python does provide an OrderedDict, which you can use if you avoid using csv.DictReader().

and it skipped the song name altogether.

This is because the file is not really a CSV file. In particular, each line begins and ends with the field separator. We can use .strip("|") to fix this.

I need all this data to be output into an array named "Music"

Then the program needs to create a dict with "Music" as a key.

I need it to have commas after each artist info. In the output I get I get

This problem is because you call json.dumps() multiple times. You should only call it once if you want a valid JSON file.

Try this:

import csv
import json
from collections import OrderedDict


def MyDictReader(fp, fieldnames):
    fp = (x.strip().strip('|').strip() for x in fp)
    reader = csv.reader(fp, delimiter="|")
    reader = ([field.strip() for field in row] for row in reader)
    dict_reader = (OrderedDict(zip(fieldnames, row)) for row in reader)
    return dict_reader

csvfile = open('music.csv', 'r')
jsonfile = open('file.json', 'w')
fieldnames = ("ID","Artist","Song", "Album")
reader = MyDictReader(csvfile, fieldnames)
json.dump({"Music": list(reader)}, jsonfile, indent=2)