Replace pandas values as index to another array

Question

Consider an array

a = np.array([5, 12, 56, 36])

and a pandas dataframe

b = pandas.DataFrame(np.array([1, 3, 0, 3, 1, 0, 2])

how does one replace the values on b by using its values as indexes for a, i.e., the intended value is:

c = pandas.DataFrame([12, 36, 5, 36, 12, 5, 56])

Can't quite figure this out.

Vaishali · Accepted Answer · 2019-04-19 19:02:32Z

2

One way is using apply,

c = b.apply(lambda x: a[x])

Or by indexing the numpy array and passing the values to DataFrame,

c = pd.DataFrame(a[b[0].values])


    0
0   12
1   36
2   5
3   36
4   12
5   5
6   56

answered Apr 19, 2019 at 18:56

Vaishali

38.5k5 gold badges62 silver badges88 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

pd.DataFrame(a[b.values])

BENY · Accepted Answer · 2019-04-19 19:12:15Z

1

Let us try something different Series.get

pd.Series(a).get(b[0])
Out[57]: 
1    12
3    36
0     5
3    36
1    12
0     5
2    56
dtype: int32

answered Apr 19, 2019 at 19:12

BENY

324k22 gold badges176 silver badges250 bronze badges

hakan.t · Accepted Answer · 2019-04-19 19:00:45Z

0

map can be used.

b.a.map({i:j for i,j in enumerate(a)})
0    12
1    36
2     5
3    36
4    12
5     5
6    56
Name: a, dtype: int64

answered Apr 19, 2019 at 19:00

hakan.t

951 silver badge9 bronze badges