Select first N columns of pandas dataframe

In this article, we will discuss different ways to select the first N columns of a dataframe in pandas.

Table of Contents

• Use iloc[] to select first N columns of pandas dataframe.
• Use [] to select first N columns of pandas dataframe.
• Use head() to select first N columns of pandas dataframe.

There are different ways to select the first N columns of a dataframe. Let’s discuss them one by one,

Use iloc[] to select first N columns of pandas dataframe

In Pandas, the Dataframe provides an attribute iloc[], to select a portion of the dataframe using position based indexing. This selected portion can be few columns or rows . We can use this attribute to select first N columns of the dataframe. For example,

N = 5
# Select first N columns
first_n_column  = df.iloc[: , :N]

We selected a portion of dataframe object, that included all rows, but only first N columns of the dataframe.

How did it work?

The syntax of dataframe.iloc[] is like,

df.iloc[row_start:row_end , col_start, col_end]

Arguments:

• row_start: The row index/position from where it should start selection. Default is 0.
• row_end: The row index/position from where it should end the selection i.e. select till row_end-1. Default is till the last row of the dataframe.
• col_start: The column index/position from where it should start selection. Default is 0.
• col_end: The column index/position from where it should end the selection i.e. select till end-1. Default is till the last column of the dataframe.

It returns a portion of the dataframe that includes rows from row_start to row_end-1 and columns from col_start to col_end-1.

To select the first N columns of the dataframe, select from column index 0 till N i.e (:N) and select all rows using default values (:),

N = 5
# Select first N columns
first_n_columns  = df.iloc[: , :N]

We provided the range to select the columns from 0 position till N, to select the first N columns, therefore it returned a dataframe. Checkout complete example to select first N columns of dataframe using iloc,

import pandas as pd

# List of Tuples
empoyees = [('Jack',  34, 11, 51, 33, 34, 77, 88) ,
            ('Riti',  31, 12, 71, 56, 55, 99, 11) ,
            ('Aadi',  16, 13, 11, 44, 55, 33, 54) ,
            ('Mark',  41, 14, 12, 78, 89, 46, 56)]

# Create a DataFrame object
df = pd.DataFrame(  empoyees)

print("Contents of the Dataframe : ")
print(df)


N = 5
# Select first N columns
first_n_columns  = df.iloc[: , :N]

print("First 5 Columns Of Dataframe : ")
print(first_n_columns)

print('Type:')
print(type(first_n_columns))

Output:

Contents of the Dataframe :
      0   1   2   3   4   5   6   7
0  Jack  34  11  51  33  34  77  88
1  Riti  31  12  71  56  55  99  11
2  Aadi  16  13  11  44  55  33  54
3  Mark  41  14  12  78  89  46  56
First 5 Columns Of Dataframe :
      0   1   2   3   4
0  Jack  34  11  51  33
1  Riti  31  12  71  56
2  Aadi  16  13  11  44
3  Mark  41  14  12  78
Type:
<class 'pandas.core.frame.DataFrame'>

We selected the first N columns of the dataframe.

Select first N columns of pandas dataframe using []

We can fetch the column names of dataframe as a sequence and then select the first N column names. Then using those column name, we can select the first N columns of dataframe using subscript operator i.e. []. For example,

print("Contents of the Dataframe : ")
print(df)

N = 5
# Select first 5 columns
first_n_columns = df[df.columns[:N]]

print("First 5 Columns Of Dataframe : ")
print(first_n_columns)

print('Type:')
print(type(first_n_columns))

Output:

Contents of the Dataframe :
      0   1   2   3   4   5   6   7
0  Jack  34  11  51  33  34  77  88
1  Riti  31  12  71  56  55  99  11
2  Aadi  16  13  11  44  55  33  54
3  Mark  41  14  12  78  89  46  56
First 5 Columns Of Dataframe :
      0   1   2   3   4
0  Jack  34  11  51  33
1  Riti  31  12  71  56
2  Aadi  16  13  11  44
3  Mark  41  14  12  78
Type:
<class 'pandas.core.frame.DataFrame'>

Use head() to select the first N columns of pandas dataframe

We can use the dataframe.T attribute to get a transposed view of the dataframe and then call the head(N) function on that view to select the first N rows i.e. the first N columns of the original dataframe. Then transpose back that dataframe object to have the column contents as a dataframe object. For example,

print("Contents of the Dataframe : ")
print(df)

N = 5

# Select first 5 columns
first_n_columns = df.T.head(N).T

print("First 5 Columns Of Dataframe : ")
print(first_n_columns)

print('Type:')
print(type(first_n_columns))

Output:

Contents of the Dataframe :
      0   1   2   3   4   5   6   7
0  Jack  34  11  51  33  34  77  88
1  Riti  31  12  71  56  55  99  11
2  Aadi  16  13  11  44  55  33  54
3  Mark  41  14  12  78  89  46  56
First 5 Columns Of Dataframe :
      0   1   2   3   4
0  Jack  34  11  51  33
1  Riti  31  12  71  56
2  Aadi  16  13  11  44
3  Mark  41  14  12  78
Type:
<class 'pandas.core.frame.DataFrame'>

It returned the first N columns of dataframe as a dataframe object.

Summary

We learned different ways to get the first N columns of a dataframe in pandas.