Sort a list of dicts according to a list of values with regex-CodePudding

I'd like to sort the keys of the list_of_dicts according to the list_months. It works fine once I remove the digits (years) from the keys of list_of_dicts, but I cannot figure out how to use the regex correctly in the lambda function to include the digits.

My code so far:

import re
list_months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
list_of_dicts = [{'Apr23': '64.401'}, {'Aug23': '56.955'}, {'Dec23': '57.453'}, {'Feb23': '90.459'}, {'Jan23': '92.731'}, {'Jul23': '56.6'}, {'Jun23': '56.509'},{'Mar23': '86.209'}, {'May23': '58.705'}, {'Nov23': '57.368'}, {'Oct23': '56.711'}, {'Sep23': '57.952'}]

r = re.compile("[a-zA-Z]{3}[0-9]{2}")
    
print(sorted(list_of_dicts, key=lambda d: [k in d for k in list_months if re.search(r, k)], reverse=True))

CodePudding user response：

No need for a regex here.

dict_months = {m:i for i, m in enumerate(list_months)}
result = sorted(list_of_dicts, key=lambda d: dict_months[next(iter(d))[:3]])
print(result)
# [{'Jan23': '92.731'}, {'Feb23': '90.459'}, {'Mar23': '86.209'}, {'Apr23': '64.401'}, {'May23': '58.705'}, {'Jun23': '56.509'}, {'Jul23': '56.6'}, {'Aug23': '56.955'}, {'Sep23': '57.952'}, {'Oct23': '56.711'}, {'Nov23': '57.368'}, {'Dec23': '57.453'}]

If you also want to take the day into account, use

def sortby(d):
    key = next(iter(d))
    return dict_months[key[:3]], int(key[3:])

result = sorted(list_of_dicts, key=sortby)

CodePudding user response：

One way using re.sub with list.index.

Note that list.index is O(n), which is quite expensive.

def get_key_loc(dic):
    k = list(dic.keys())[0]
    return list_months.index(re.sub("\d ", "", k))

sorted(list_of_dicts, key=get_key_loc)

Output:

[{'Jan23': '92.731'},
 {'Feb23': '90.459'},
 {'Mar23': '86.209'},
 {'Apr23': '64.401'},
 {'May23': '58.705'},
 {'Jun23': '56.509'},
 {'Jul23': '56.6'},
 {'Aug23': '56.955'},
 {'Sep23': '57.952'},
 {'Oct23': '56.711'},
 {'Nov23': '57.368'},
 {'Dec23': '57.453'}]

CodePudding user response：

You don't really need the dict_month, use python's batteries included:

from datetime import datetime

sorted(list_of_dicts, key=lambda x: datetime.strptime(next(iter(x)), '%b%d'))

or:

import dateutil.parser

sorted(list_of_dicts, key=lambda x: dateutil.parser.parse(next(iter(x))))

output:

[{'Jan23': '92.731'},
 {'Feb23': '90.459'},
 {'Mar23': '86.209'},
 {'Apr23': '64.401'},
 {'May23': '58.705'},
 {'Jun23': '56.509'},
 {'Jul23': '56.6'},
 {'Aug23': '56.955'},
 {'Sep23': '57.952'},
 {'Oct23': '56.711'},
 {'Nov23': '57.368'},
 {'Dec23': '57.453'}]