Home How to efficiently remove elements from dicts that have certain value patterns?

Questions

How to efficiently remove elements from dicts that have certain value patterns?

March 17, 2022

For example, in dict1 the keys 1, 2, 3 all have the same value ‘a’, but the keys 3 and 5 have different values, ‘b’ and ‘d’. What I want is:

If N keys have the same value and N >=3, then I want to remove all other elements from the dict and only keep those N key values, which means ‘b’ & ‘d’ have to be removed from the dict.

The following code works, but it seems very verbose. Is there a better way to do this?

from collections import defaultdict

dict1 = {1:'a', 2:'a', '3':'b', '4': 'a', '5':'d'}
l1 = [1, 2, 3, 4, 5]

dict2 = defaultdict(list)
for k, v in dict1.items():
    dict2[v].append(k)

to_be_removed = []
is_to_be_removed = False
for k, values in dict2.items():
    majority = len(values)
    if majority>=3:
        is_to_be_removed = True
    else:
        to_be_removed.extend(values)

if is_to_be_removed:
    for d in to_be_removed:
        del dict1[d]

print(f'New dict: {dict1}')

>Solution :

You can use collections.Counter to get the frequency of every value, then use a dictionary comprehension to retain only the keys that have the desired corresponding value:

from collections import Counter

dict1 = {1:'a', 2:'a', '3':'b', '4': 'a', '5':'d'}
ctr = Counter(dict1.values())

result = {key: value for key, value in dict1.items() if ctr[value] >= 3}
print(result)

This outputs: