Starting from dataframe df, I would like to apply dictionary dict as translation table and create another column with translated results:
input:
df=pd.DataFrame({'A': 'foo bar one123 bar foo one324 foo 0'.split(),
'B': 'one546 one765 twosde three twowef two234 onedfr three'.split(),
'C': 'lol'})
output:
A B C
0 foo one546 lol
1 bar one765 lol
2 one123 twosde lol
3 bar three lol
4 foo twowef lol
5 one324 two234 lol
6 foo onedfr lol
7 0 three lol
replacement dictionary:
dict={'foo':['blue'],'bar':['red'],'one123':['purple','green']}
For each occurrance of a key of dict in column "A", I would like to assign value of dict to a new column "D". Notice that some value of dict are list with one or two elements.
for key,value in dict.items():
df.loc[df['A']==key,'D']=[value]
I tried with and without square bracket on value, so value and [value] but the result is always:
ValueError: Must have equal len keys and value when setting with an ndarray
I noticed that if I put df.loc[df['A']==key,'D']=value, so without square brackets around value, I obtain the error before the first substitution, while if I use square brackets, the error appears before the last element ['purple','green'], so the list with two elements.
How can I prevent pandas from trying to unpack the list and just use value as it is? Thank you for the kind answer