按由数字和字母组成的键对字典进行排序

Question

I have a dict that contains data about different quarters (of years):

"data": {
        "Q3/2016": 86,
        "Q1/2016": 85,
        "Q4/2016": 69,
        "Q2/2016": 69,
        "Q2/2017": 82,
        "Q1/2017": 66,
    },

How can I sort this so it looks like this

"data": {
        "Q1/2016": 85,
        "Q2/2016": 69,
        "Q3/2016": 86,
        "Q4/2016": 69,
        "Q1/2017": 66,
        "Q2/2017": 82,
    },

Answer 1

Dictionaries cannot be ordered in-place since they have no order, but since Python 3.7 they remember insertion order so we can build an "ordered" dictionary by sorting key-value pairs from your unordered dictionary.

So the actual important part is the key we pass as an argument to sorted . In the following code, the key is a lambda function that takes each key, split it at / and reverses the resulting tuple, so that keys are sorted first by year, then by quarter:

d = {"data": { 
    "Q3/2016": 86, 
    "Q1/2016": 85, 
    "Q4/2016": 69, 
    "Q2/2016": 69, 
    "Q2/2017": 82, 
    "Q1/2017": 66, 
    }
}
d['data'] = {k: v for k, v in sorted(d['data'].items(), key=lambda x: x[0].split('/')[::-1])}
print(d)

output:

{'data': {
     'Q1/2016': 85,
     'Q2/2016': 69,
     'Q3/2016': 86,
     'Q4/2016': 69,
     'Q1/2017': 66,
     'Q2/2017': 82}
}

Answer 2

You should copy this dict into an ordered dict like

In [22]: from collections import OrderedDict
In [23]: data = {
    ...:         "Q3/2016": 86,
    ...:         "Q1/2016": 85,
    ...:         "Q4/2016": 69,
    ...:         "Q2/2016": 69,
    ...:         "Q2/2017": 82,
    ...:         "Q1/2017": 66,
    ...:     }
In [40]: def sort_key(key_val):
    ...:     key, val = key_val
    ...:     key = key.split("/")
    ...:     return int(key[1]), key[0]
    ...:
# add a custom sort key
In [24]: data = OrderedDict(sorted(data.items(), key=sort_key))

In [25]: data
Out[25]:
OrderedDict([('Q1/2016', 85),
             ('Q2/2016', 69),
             ('Q3/2016', 86),
             ('Q4/2016', 69),
             ('Q1/2017', 66),
             ('Q2/2017', 82)])

It seems to maintain the order while json serialisation too which is nice

In [27]: print(json.dumps(data, indent=4))
{
    "Q1/2016": 85,
    "Q2/2016": 69,
    "Q3/2016": 86,
    "Q4/2016": 69,
    "Q1/2017": 66,
    "Q2/2017": 82
}