CODEWITHSUNDEEP
pythonpandassyntaxapply
Ewdlam
Ewdlam

Reputation: 935

Pandas apply, 'float' object is not subscriptable

I have the following dataframe :

Y = list(range(5))
Z = np.full(5, np.nan)
df = pd.DataFrame(dict(ColY = Y, ColZ = Z))
print(df)
   ColY  ColZ
0     0   NaN
1     1   NaN
2     2   NaN
3     3   NaN
4     4   NaN

And this dictionary :

Dict = {
    0 : 1,
    1 : 2,
    2 : 3,
    3 : 2,
    4 : 1
}

I would like to fill ColZ with "ok" if corresponding value of ColY through Dict is 2. Consequently, I would like the following dataframe :

   ColY ColZ
0     0  NaN
1     1   ok
2     2  NaN
3     3   ok
4     4  NaN

I tried this script:

df['ColZ'] = df['ColZ'].apply(lambda x : "ok" if Dict[x['ColY']] == 2 else Dict[x['ColY']])

I have this error :

TypeError: 'float' object is not subscriptable

Do you know why ?

Upvotes: 1

Views: 1515

Answers (1)

jezrael
jezrael

Reputation: 862641

Use numpy.where with Series.map for new Series for compare by Series.eq (==):

df['ColZ'] = np.where(df['ColY'].map(Dict).eq(2), 'ok', np.nan)
print(df)
   ColY ColZ
0     0  nan
1     1   ok
2     2  nan
3     3   ok
4     4  nan

Detail:

print(df['ColY'].map(Dict))
0    1
1    2
2    3
3    2
4    1
Name: ColY, dtype: int64

Your solution should be changed with .get for return some default value, here np.nan if no match:

df['ColZ'] = df['ColY'].apply(lambda x : "ok" if Dict.get(x, np.nan) == 2 else np.nan)

EDIT: For set working with df['ColZ'] values use:

Y = list(range(5))
Z = list('abcde')
df = pd.DataFrame(dict(ColY = Y, ColZ = Z))
print(df)
Dict = {
    0 : 1,
    1 : 2,
    2 : 3,
    3 : 2,
    4 : 1
}

df['ColZ1'] = np.where(df['ColY'].map(Dict).eq(2), 'ok', df['ColZ'])
df['ColZ2'] = df.apply(lambda x : "ok" if Dict.get(x['ColY'], np.nan) == 2 
                                       else x['ColZ'], axis=1)
print (df)
   ColY ColZ ColZ1 ColZ2
0     0    a     a     a
1     1    b    ok    ok
2     2    c     c     c
3     3    d    ok    ok
4     4    e     e     e

Upvotes: 2

Related Questions