let's suppose that my dataset looks like this:

I have some key values being defined:

school_type_key={'State School':0, 'Private National School':1, 'Private International School':2}
specialization_key={'STEM':0,'Humanities':1,'Arts':2, 'Life Sciences':3, 'General':4, 'Other':5}

I need to write a function that:

1. takes a column
2. makes a list of unique value
3. sorts this list based on defined keys
4. returns a list

Here is what I tried:

def sort_by_keys(column):
    unique_values=df[column].unique()
    unique_values=list(unique_values)
    key=column.replace(' ','_').lower() '_key'

    #I need key so that if i pass 'School Type' column, it is sorted by 'school_type_key'

    unique_values.sort(key=key.get, reverse=True)
    return unique_values

sort_by_keys('School Type')

However, it returns an error because key is a string and 'get' can't be applied towards a string. How can I solve this issue?

Expected output:

output=sort_by_keys('School Type')
print(output)
['State School', 'Private National School', 'Private International School']

output=sort_by_keys('Specialization')
print(output)
['STEM', 'Humanities','Arts','General']

CodePudding user response：

You can use:

MAPPING = {
    'School Type': school_type_key,
    'Specialization': specialization_key
}

def sort_by_keys(col):
    return (df[col].sort_values(key=lambda x: x.map(MAPPING[col]))
                   .unique().tolist())

Output:

>>> sort_by_keys('School Type')
['State School', 'Private National School', 'Private International School']

>>> sort_by_keys('Specialization')
['STEM', 'Humanities', 'Arts', 'General']