CODEWITHSUNDEEP
pythonpandas
Kustekjé Meklootn
Kustekjé Meklootn

Reputation: 1328

Get column by number in Pandas

What's the difference between:

Maand['P_Sanyo_Gesloten']
Out[119]: 
Time
2012-08-01 00:00:11    0
2012-08-01 00:05:10    0
2012-08-01 00:10:11    0
2012-08-01 00:20:10    0
2012-08-01 00:25:10    0
2012-08-01 00:30:09    0
2012-08-01 00:40:10    0
2012-08-01 00:50:09    0
2012-08-01 01:05:10    0
2012-08-01 01:10:10    0
2012-08-01 01:15:10    0
2012-08-01 01:25:10    0
2012-08-01 01:30:10    0
2012-08-01 01:35:09    0
2012-08-01 01:40:10    0
...
2012-08-30 22:35:09    0
2012-08-30 22:45:10    0
2012-08-30 22:50:09    0
2012-08-30 22:55:10    0
2012-08-30 23:00:09    0
2012-08-30 23:05:10    0
2012-08-30 23:10:09    0
2012-08-30 23:15:10    0
2012-08-30 23:20:09    0
2012-08-30 23:25:10    0
2012-08-30 23:35:09    0
2012-08-30 23:40:10    0
2012-08-30 23:45:09    0
2012-08-30 23:50:10    0
2012-08-30 23:55:11    0
Name: P_Sanyo_Gesloten, Length: 7413, dtype: int64

And

Maand[[1]]
Out[120]: 
&ltclass 'pandas.core.frame.DataFrame'&gt
DatetimeIndex: 7413 entries, 2012-08-01 00:00:11 to 2012-08-30 23:55:11
Data columns (total 1 columns):
P_Sanyo_Gesloten    7413  non-null values
dtypes: int64(1)

How can I get column by its index number? And not by an index string?

Upvotes: 52

Views: 216917

Answers (6)

Asclepius
Asclepius

Reputation: 63453

To formalize the comment by Jeff into an answer, this is the simplest I know that works. It is simpler than the prior answer by Andy H which uses a list.

my_col = df.iloc[:, column_index]

Use 0 for the first column, 1 for the second column, and so on. For example:

first_col = df.iloc[:, 0]

As a bonus, if the goal is to iterate over the columns, df.items() is sufficient.

Upvotes: 3

rachwa
rachwa

Reputation: 2310

You can also use take to get any column(s) by position:

In [2]: df = pd.DataFrame([[1, 2], [3, 4]], columns=['a', 'b'])

In [3]: df
Out[3]: 
   a  b
0  1  2
1  3  4

In [4]: df.take([1], axis=1)
Out[4]: 
   b
0  2
1  4

Upvotes: 1

ListenSoftware Louise Ai Agent
ListenSoftware Louise Ai Agent

Reputation: 4253

another way to access a column by number is to use a mapping dictionary where the key is the column name and the value is the column number

dates = pd.date_range('1/1/2000', periods=8)

df = pd.DataFrame(np.random.randn(8, 4),
   index=dates, columns=['A', 'B', 'C', 'D'])
print(df)
dct={'A':0,'B':1,'C':2,'D':3}
columns=df.columns

print(df.iloc[:,dct['D']])

Upvotes: 1

Sam
Sam

Reputation: 12049

Another way is to select a column with the columns array:

In [5]: df = pd.DataFrame([[1,2], [3,4]], columns=['a', 'b'])

In [6]: df
Out[6]: 
   a  b
0  1  2
1  3  4

In [7]: df[df.columns[0]]
Out[7]: 
0    1
1    3
Name: a, dtype: int64

Upvotes: 23

Andy Hayden
Andy Hayden

Reputation: 375685

One is a column (aka Series), while the other is a DataFrame:

In [1]: df = pd.DataFrame([[1,2], [3,4]], columns=['a', 'b'])

In [2]: df
Out[2]:
   a  b
0  1  2
1  3  4

The column 'b' (aka Series):

In [3]: df['b']
Out[3]:
0    2
1    4
Name: b, dtype: int64

The subdataframe with columns (position) in [1]:

In [4]: df[[1]]
Out[4]:
   b
0  2
1  4

Note: it's preferable (and less ambiguous) to specify whether you're talking about the column name e.g. ['b'] or the integer location, since sometimes you can have columns named as integers:

In [5]: df.iloc[:, [1]]
Out[5]:
   b
0  2
1  4

In [6]: df.loc[:, ['b']]
Out[6]:
   b
0  2
1  4

In [7]: df.loc[:, 'b']
Out[7]:
0    2
1    4
Name: b, dtype: int64

Upvotes: 59

pandita
pandita

Reputation: 4989

The following is taken from http://pandas.pydata.org/pandas-docs/dev/indexing.html. There are a few more examples... you have to scroll down a little

In [816]: df1

           0         2         4         6
0   0.569605  0.875906 -2.211372  0.974466
2  -2.006747 -0.410001 -0.078638  0.545952
4  -1.219217 -1.226825  0.769804 -1.281247
6  -0.727707 -0.121306 -0.097883  0.695775
8   0.341734  0.959726 -1.110336 -0.619976
10  0.149748 -0.732339  0.687738  0.176444

Select via integer slicing

In [817]: df1.iloc[:3]

          0         2         4         6
0  0.569605  0.875906 -2.211372  0.974466
2 -2.006747 -0.410001 -0.078638  0.545952
4 -1.219217 -1.226825  0.769804 -1.281247

In [818]: df1.iloc[1:5,2:4]

          4         6
2 -0.078638  0.545952
4  0.769804 -1.281247
6 -0.097883  0.695775
8 -1.110336 -0.619976

Select via integer list

In [819]: df1.iloc[[1,3,5],[1,3]]

           2         6
2  -0.410001  0.545952
6  -0.121306  0.695775
10 -0.732339  0.176444

Upvotes: 18

Related Questions