Data Selection with loc[], iloc[] and at[] for DataFrames in Pandas

The loc, iloc and at (and ix, which is deprecated) methods provide flexible ways to access and manipulate data in a DataFrames and are useful for data exploration, data cleaning, and data manipulation tasks in data analysis and machine learning workflows.

Below you will find a summary of how to use these methods with short explanation and code examples.

For all examples we use the same, simple DataFrame:

import pandas as pd

df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6], 'C': [7, 8, 9]}, index=['a', 'b', 'c'])

#    A  B  C
# a  1  4  7
# b  2  5  8
# c  3  6  9

loc

The loc function is used to index and select data using labels ("label-based"): You have to specify row and column labels for the data selection.

# Select a single row by label
df.loc['a']
# A    1
# B    4
# C    7

# Select multiple rows by label
df.loc[['a', 'b']]
#    A  B  C
# a  1  4  7
# b  2  5  8

# Select a single value by row and column label
df.loc['a', 'A']
# 1

# Select multiple values by row and column labels
df.loc[['a', 'b'], ['A', 'B']]
#   A  B
# a  1  4
# b  2  5

iloc

The iloc function is used to index and select data using integers. It is "integer-based" and is used to select data by specifying rows and columns based on their row and column indices.

# Select a single row by index
df.iloc[0]
# A    1
# B    4
# C    7

# Select multiple rows by index
df.iloc[[0, 1]]
#    A  B  C
# a  1  4  7
# b  2  5  8

# Select a single value by row and column index
df.iloc[0, 0]
# 1

# Select multiple values by row and column indices
df.iloc[[0, 1], [0, 1]]
#    A  B
# a  1  4
# b  2  5

at

The at function is used to access a single value at a given row and column label. It is faster than using the loc function to access a single value.

# Access a single value by row and column label
df.at['a', 'A']
# 1

ix

The ix function is a deprecated indexer that used to combine the functionality of loc and iloc for label-based and integer-based indexing.

# Select a single row by label or index
print(df.ix[0])

# Select multiple rows by label or index
print(df.ix[['a', 'b']])

# Access a single value by row and column label or index
print(df.ix[0, 'A'])

# Access multiple values by row and column label or index
print(df.ix[['a', 'b'], ['A', 'B']])

Summary

• if you want to index and select data based on row and column labels, use loc
• If you want to index and select data based on row and column indices, use iloc.
• If you want to access a single value in the DataFrame, use at