Redefining the Index in a Pandas DataFrame object

Question

I am trying to re-index a pandas DataFrame object, like so,

From:
            a   b   c
        0   1   2   3
        1  10  11  12
        2  20  21  22

To :
           b   c
       1   2   3
      10  11  12
      20  21  22

I am going about this as shown below and am getting the wrong answer. Any clues on how to do this?

>>> col = ['a','b','c']
>>> data = DataFrame([[1,2,3],[10,11,12],[20,21,22]],columns=col)
>>> data
    a   b   c
0   1   2   3
1  10  11  12
2  20  21  22
>>> idx2 = data.a.values
>>> idx2
array([ 1, 10, 20], dtype=int64)
>>> data2 = DataFrame(data,index=idx2,columns=col[1:])
>>> data2
     b   c
1   11  12
10 NaN NaN
20 NaN NaN

Any idea why this is happening?

because you are using the 1st DF to build the second, this will only get the rows where idx2 intersect the data.index, ie. row 1 — RuiDC, Jun 27 '12 at 12:56

score 191 · Accepted Answer · answered May 05 '12 at 02:44

191

Why don't you simply use set_index method?

In : col = ['a','b','c']

In : data = DataFrame([[1,2,3],[10,11,12],[20,21,22]],columns=col)

In : data
Out:
    a   b   c
0   1   2   3
1  10  11  12
2  20  21  22

In : data2 = data.set_index('a')

In : data2
Out:
     b   c
a
1    2   3
10  11  12
20  21  22

answered May 05 '12 at 02:44

Avaris

35,883
7
81
72

18

In order to remove the index name, as in the original example: data2.index.name = None – Daniele Nov 03 '14 at 18:31

score 5 · Answer 2 · answered May 10 '17 at 12:39

5

If you don't want 'a' in the index

In :

col = ['a','b','c']

data = DataFrame([[1,2,3],[10,11,12],[20,21,22]],columns=col)

data

Out:

    a   b   c
0   1   2   3
1  10  11  12
2  20  21  22

In :

data2 = data.set_index('a')

Out:

In :

data2.index.name = None

Out:

answered May 10 '17 at 12:39

Aakash Saxena

199
3
8

1

The second Out is not correct. There is still an unnamed column on the left containing 0, 1, 2. – Yster Jun 20 '17 at 15:44

score 0 · Answer 3 · answered Mar 24 '21 at 09:17

To avoid index name with a single-liner, you can use set_index('a') with rename_axis(None)

In [8]: data.set_index('a').rename_axis(None)
Out[8]: 
     b   c
1    2   3
10  11  12
20  21  22

Details

In [9]: data
Out[9]: 
    a   b   c
0   1   2   3
1  10  11  12
2  20  21  22

Redefining the Index in a Pandas DataFrame object

3 Answers3

Linked

Related