Selecting columns | The School of Data

Introduction to Python for Data Analysis Selecting columns

goals In this section, we'll learn how to:

Select columns using pandas
Select columns based on their positions
Select a range of columns
Drop columns
Select columns based on their data type

A step by step approach using method chaining

In Python with pandas, we can use method chaining to perform operations step by step. This is similar to the pipe operator in R, but we use the dot (.) to chain methods in Python.

Selecting columns with column names

In pandas, we can select columns from a DataFrame using square brackets [] or the loc accessor.

Syntax

You can read the following syntax as “take the dataset, and then select column1 and column2”.

 df = dataset[['column1', 'column2']]
print(df)
# or
df = dataset.loc[:, ['column1', 'column2']]
print(df)

Example

In our flowers dataset, we have these columns: name , height , season , sunlight , and growth .

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

If we want to select the name and height columns, we can do this:

 df = flowers[['name', 'height']]
print(df)
# or
df = flowers.loc[:, ['name', 'height']]
print(df)

This code will select the name and height columns from the flowers dataset and print the df.

Exercise 2.1

Run the code below to see the output of selecting the name and height columns from the flowers dataset.

Exercise 2.2

Select the season and sunlight columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Selecting columns using column positions

We can also select columns using their positions with the iloc accessor.

Example

In the flowers dataset, we can select the first and third columns using their positions.

 df = flowers.iloc[:, [0, 2]]
print(df)

This code will select the first and third columns from the flowers dataset and print the df.

Exercise 2.3

Run the code below to see the output of selecting the first and third columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Exercise 2.4

Try selecting the second and fourth columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Selecting a range of columns

We can select a range of columns using slice notation.

Example

 df = flowers.iloc[:, 0:3]
print(df)

This code will select the first three columns from the flowers dataset and print the df.

Exercise 2.5

Run the code below to see the output of selecting the first three columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Exercise 2.6

Try selecting the columns from the second to the fourth column. Hint: Use 1:4 in the slice notation.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Dropping columns

If you want to drop columns, you can use the drop method.

Example

 df = flowers.drop('name', axis=1)
print(df)

This code will drop the name column from the flowers dataset and print the df.

Selecting Columns based on Data Type

You can select columns based on their data type using the select_dtypes method.

Selecting numeric columns

 df = dataset.select_dtypes(include=['number'])
print(df)

Selecting object (string) columns

 df = dataset.select_dtypes(include=['object'])
print(df)

Example

In the flowers dataset, we can select only the numeric columns:

 df = flowers.select_dtypes(include=['number'])
print(df)

This code will select only the height and sunlight columns from the flowers dataset and print the df.

Exercise 2.7

Run the code below to see the output of selecting only the numeric columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Exercise 2.8

Select only the object (string) columns from the flowers dataset.

name	height	season	sunlight	growth
Poppy	75	Spring	8.3	fast
Rose	150	Summer	6.4	slow
Zinnia	60	Summer	8.7	fast
Peony	90	Spring	7.2	slow

Quiz

summary In this section, we've learned how to:

Select columns using pandas
Select columns based on their positions
Select a range of columns
Drop columns
Select columns based on their data type

Previous What is Python?

Go to Next SectionFiltering Rows →

On this page

- A step by step approach using method chaining
- Selecting columns with column names
- Selecting columns using column positions
- Selecting a range of columns
- Dropping columns
- Selecting Columns based on Data Type