How Can I Efficiently Change Multiple Column Data Types in Pandas?

Changing Column Types in Pandas for Multiple Columns

To convert multiple columns in a DataFrame to specific types, consider using the following methods:

Using Pandas' to_numeric()

This method can safely convert non-numeric types, such as strings, into integers or floating-point numbers as appropriate. For example:

import pandas as pd

table = [
    ['a', '1.2', '4.2'],
    ['b', '70', '0.03'],
    ['x', '5', '0'],
]

df = pd.DataFrame(table)

# Convert columns 2 and 3 to floats
df[['Column2', 'Column3']] = df[['Column2', 'Column3']].apply(pd.to_numeric)

Using Pandas' astype()

This method allows explicit conversion to a specified dtype. For example:

df[['Column2', 'Column3']] = df[['Column2', 'Column3']].astype(float)

Conversion Methods Overview

The choice of method depends on the specific requirements and data structure:

to_numeric(): Ideal for reliable conversion from non-numeric values to numeric types.

astype(): Explicit and flexible conversion to any desired dtype.

infer_objects(): Introduced in pandas 0.21.0, specifically for converting object columns to a more specific type.

convert_dtypes(): Part of pandas version 1.0 and above, automatically converts columns to the "best possible" type that supports pandas' NA missing value.

The above is the detailed content of How Can I Efficiently Change Multiple Column Data Types in Pandas?. For more information, please follow other related articles on the PHP Chinese website!