Altcademy - a Forbes magazine logo Best Coding Bootcamp 2023

How to use loc in Pandas

Understanding the Basics of Pandas

Before we dive into the specifics of using loc in Pandas, it's important to have a basic understanding of what Pandas is. Pandas is a powerful data manipulation library in Python that makes it easy to work with structured data, like tables. It provides data structures and functions that make it simple to perform complex operations on datasets.

The DataFrame: Your Data's New Home

At the heart of Pandas is the DataFrame—a two-dimensional, size-mutable, and potentially heterogeneous tabular data structure with labeled axes (rows and columns). You can think of a DataFrame as a spreadsheet or a SQL table. It's a convenient way to store and manipulate data.

The Power of loc: Accessing Your Data

The loc attribute is one of the many ways provided by Pandas to select and manipulate data. The word "loc" stands for "location," and it's used to access a group of rows and columns by labels or a boolean array. You can think of loc as a sophisticated version of indexing you might have seen in Python lists, but with much more power.

The Syntax of loc

The basic syntax of loc is straightforward:

dataframe.loc[<row_labels>, <column_labels>]

Here, <row_labels> and <column_labels> can be:

  • Single labels
  • Lists of labels
  • A slice object with labels
  • A boolean array

Selecting Rows with loc

To select rows using loc, you pass the index labels of the rows you're interested in. Let's say we have a DataFrame named df with some data about fruits:

import pandas as pd

data = {
    'fruit': ['apple', 'banana', 'cherry', 'date'],
    'color': ['red', 'yellow', 'red', 'brown'],
    'weight': [180, 120, 10, 5]

df = pd.DataFrame(data)
df.index = ['a', 'b', 'c', 'd']  # Setting custom row labels

If we want to select the row with the label 'b', we use loc like this:


This will output the information about bananas in our DataFrame.

Selecting Columns with loc

Similarly, if you want to select a specific column, you can do so by specifying the column label. Let's say we want to select the 'color' column:

print(df.loc[:, 'color'])

The colon : before the comma indicates that we want all rows, and 'color' specifies the column we are interested in.

Selecting Both Rows and Columns

loc also allows you to select both rows and columns simultaneously. Let's say we want to know the color and weight of the cherry:

print(df.loc['c', ['color', 'weight']])

This will give us the color and weight of the cherry, by selecting row 'c' and the columns 'color' and 'weight'.

Using Slices with loc

Just like with Python lists, you can use slice notation with loc to select a range of rows or columns. For example, to select all fruits from banana to date:


Remember that, unlike standard Python slicing, the end label in Pandas' slices is inclusive.

Conditional Selection with loc

One of the most powerful features of loc is the ability to perform conditional selections. Suppose you want to find all fruits that are red. You can do this by:

print(df.loc[df['color'] == 'red'])

Here, df['color'] == 'red' creates a boolean array that loc uses to select rows where the condition is True.

Setting Values with loc

loc isn't just for selecting data; you can also use it to set values. If we want to change the weight of the apple to 200 grams, we would do:

df.loc['a', 'weight'] = 200

After executing this code, the weight of the apple in our DataFrame will be updated to 200.

Avoiding Common Mistakes

When using loc, it's important to remember that it works with labels, not integer positions. If you try to use loc with an integer index when your DataFrame has custom labels, you'll run into errors. In such cases, you'll want to use iloc, which is designed for integer-location based indexing.

Intuition and Analogies

Think of the DataFrame as a big office cabinet with many drawers (rows) and sections (columns). The loc is like telling a coworker, "Please fetch the contents of the top drawer, second section." You're using specific labels, not the numerical order of the drawers or sections.

Advanced Usage: Slicing with Labels and Boolean Arrays

What if you want to select all fruits that weigh more than 100 grams and only show their color? You can combine slicing with a boolean array:

print(df.loc[df['weight'] > 100, 'color'])

This will display the colors of all fruits that weigh more than 100 grams.

Conclusion: Unlocking Data Potential with loc

Mastering the use of loc in Pandas can feel like learning the combinations to a powerful safe filled with your data's secrets. With it, you can unlock precise subsets of data, peer into the intricate details of your dataset, and even rearrange the contents to your liking. Whether you're a budding data analyst or a seasoned programmer, understanding how to use loc effectively can greatly enhance your data manipulation skills, allowing you to handle your data with both precision and ease. So next time you're faced with a daunting dataset, remember that loc is your trusty key to unlocking the information you need just when you need it.