python loc vs iloc

Guri Bee

3 min read

·

17-10-2024

1

0

Share

Unlocking the Power of Data Selection in Python: Loc vs Iloc Explained

When working with data in Python, particularly with Pandas DataFrames, the ability to select specific rows and columns is crucial. Two fundamental methods, .loc and .iloc, provide powerful ways to access and manipulate data, but their subtle differences can be confusing. This article aims to demystify the distinction between .loc and .iloc, empowering you to confidently navigate your datasets.

Understanding the Basics

Imagine your Pandas DataFrame as a spreadsheet, where each row represents a record and each column represents a feature. Both .loc and .iloc help you access specific cells within this spreadsheet, but they operate differently.

• .loc (Label-based Selection): This method uses labels (row and column names) to identify and select data. It's like pointing to a cell by saying "I want the value in the row labeled 'Customer ID' and the column labeled 'Purchase Date'".

• .iloc (Integer-based Selection): In contrast, .iloc relies on numerical indices (positions within the DataFrame) for selection. This is akin to saying "Give me the value in the third row and the second column".

Practical Examples

Let's illustrate with a simple example:

import pandas as pd

data = {'Name': ['Alice', 'Bob', 'Charlie', 'David'],
        'Age': [25, 30, 28, 32],
        'City': ['New York', 'London', 'Paris', 'Tokyo']}

df = pd.DataFrame(data)
print(df)

This code creates a DataFrame named df containing information about four individuals.

1. Using .loc

• Selecting a single cell:

  print(df.loc['Bob', 'Age'])  # Output: 30
  

  This retrieves the value at the intersection of the row labeled 'Bob' and the column labeled 'Age'.

• Selecting multiple rows:

  print(df.loc[['Alice', 'Charlie'], :]) 
  

  This selects the rows labeled 'Alice' and 'Charlie' and all columns.

• Selecting a range of rows:

  print(df.loc['Bob':'David', 'City']) 
  

  This selects rows from 'Bob' to 'David' (inclusive) and the column 'City'.

2. Using .iloc

• Selecting a single cell:

  print(df.iloc[2, 1])  # Output: 28
  

  This retrieves the value at the intersection of the 3rd row (index 2) and the 2nd column (index 1).

• Selecting multiple rows:

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

  This selects the rows at indices 0 and 2 (the 1st and 3rd rows).

• Selecting a range of rows:

  print(df.iloc[1:3, 0]) 
  

  This selects rows from index 1 to 3 (exclusive of 3) and the 1st column (index 0).

Key Differences and Considerations

• Flexibility: .loc allows for more flexible selections by using labels, which can be string-based or even multi-level indices.

• Performance: .iloc can be slightly faster for large datasets as it relies on direct numerical indexing.

• Consistency: Using .loc promotes consistency by focusing on meaningful labels, making your code more readable and less prone to errors due to index shifts.

When to Use Which Method

• .loc is your go-to choice when:

  • You know the labels of the rows and columns you want to access.
  • You want to select data based on conditions or boolean indexing.
  • You need a clear and descriptive way to select data.

• .iloc is preferred when:

  • You need to access data using positional indices.
  • You're performing operations that rely on numeric positions, like slicing or selecting a specific number of rows.

Further Exploration

Beyond .loc and .iloc, Pandas offers a powerful selection toolkit. You can explore features like df.at for single-cell access, df.iat for integer-based single-cell access, and Boolean indexing for selecting based on conditions.

Conclusion

Understanding the differences between .loc and .iloc is fundamental to efficient and effective data analysis in Python. By mastering these methods, you'll unlock the full power of Pandas DataFrames and gain a powerful advantage in your data manipulation tasks.

• Original GitHub Source: https://github.com/pandas-dev/pandas/issues/20784 (Thank you to the Pandas developers for their contributions to this library!)