df-squeezer 0.0.5

A small package to optimize memory usage of Pandas DataFrames

DF-Squeezer

DF-Squeezer is a Python package designed to help reduce the memory usage of pandas DataFrames. By analyzing the data types of DataFrame columns, DF-Squeezer suggests and can automatically apply data type conversions that are more memory-efficient without losing information.

Features

• Analysis of DataFrame Columns: Identifies columns that can be converted to more memory-efficient data types.
• Memory Usage Reporting: Calculates and reports the potential memory savings from suggested conversions.
• Automatic Conversion Application: Optionally applies suggested conversions directly to the DataFrame.

Installation

To install DF-Squeezer, run the following command in your terminal:

pip install df-squeezer

Ensure you have pandas installed in your environment as it is a required dependency.

Usage

from df_squeezer import df_squeezer
import pandas as pd

# Sample DataFrame
df = pd.DataFrame({
    'a': range(1000),
    'b': [float(i) for i in range(1000)],
    'c': ['category' + str(i % 3) for i in range(1000)]
})

# Use df_squeeze to analyze and report potential dtype conversions
optimized_df = df_squeezer(df, report=True, edit=False)

# To directly apply the suggested conversions
optimized_df = df_squeezer(df, report=True, edit=True)

Parameters

df (pd.DataFrame): The input DataFrame to analyze and optimize. report (bool, optional): If True (default), prints suggested changes for data type conversions along with potential memory savings. edit (bool, optional): If False (default), the function only reports potential optimizations without applying them. If True, the function applies the suggested data type conversions directly to the DataFrame. Returns pd.DataFrame: The original DataFrame with applied changes if edit=True. Otherwise, returns the original DataFrame unmodified. Contributing Contributions to DF-Squeezer are welcome! Please refer to the contribution guidelines for more information.

License This project is licensed under the MIT License - see the LICENSE file for details.