Python: Iterate over values in a series and replace with dictionary values when key matches series value

user_12209200 0
Viewed 35

I'm trying to iterate over a column in a dataframe and when the value matches a key from my dictionary it should then replace the value in another column with the value of the matching key.

    df = pd.DataFrame({'id': ['123', '456', '789'], 'Full': ['Yes', 'No', 'Yes'], 'Cat':['','','']})
    cats = {'123':'A', '456':'B', '789':'C'}
    for val in df.id:
        for key, cat in cats.items():
            if key == val:
                df.Cat.loc[(df.Full == 'Yes')] = cat
    df
        id  Full Cat
    0   123 Yes  C
    1   456 No  
    2   789 Yes  C

I would expect id 123 to have a Cat of 'A' but instead it only returns 'C'

Can anyone explain to me why the it isn't iterating over the keys in dictionary?

pandas dictionary python
2 Answers

You can use Series.replace and pass the dictionary, and assign the result to Cat column:

>>> df['Cat'] = df.id.replace(cats)
#output:

    id Full Cat
0  123  Yes   A
1  456   No   B
2  789  Yes   C

Or, if you intend to replace in only the rows with Full as Yes one way is to simply apply a function on axis=1 then implement the logic for each rows:

>>> df['Cat'] = df.apply(lambda x: cats.get(x.id, '') if x.Full == 'Yes' else '',
                     axis=1)

    id Full Cat
0  123  Yes   A
1  456   No    
2  789  Yes   C
user_91363480

For filtered values in column use dict.get:

mask = df.Full == 'Yes'
df.loc[mask, 'Cat'] = df.loc[mask, 'id'].apply(lambda x: cats.get(x, ''))
print (df)
    id Full Cat
0  123  Yes   A
1  456   No    
2  789  Yes   C

If no match in dict is possible create Nones use:

cats = {'123':'A', '456':'B', '7890':'C'}

mask = df.Full == 'Yes'
df.loc[mask, 'Cat'] = df.loc[mask, 'id'].apply(cats.get)
print (df)
    id Full   Cat
0  123  Yes     A
1  456   No      
2  789  Yes  None
user_29010020
Related