panars: Polars with Pandas-like Interface

Panars is a powerful wrapper that brings the familiar Pandas API to Polars, combining the best of both worlds: Polars' speed and efficiency with Pandas' user-friendly interface.

Key Features

• Pandas-like API: Use Polars with syntax you already know from Pandas.
• High Performance: Leverage Polars' speed while writing Pandas-style code.
• Easy Migration: Seamlessly transition existing Pandas code to Polars.
• Best of Both Worlds: Combine Pandas' ease of use with Polars' efficiency.

Installation

pip install panars

Quick Start

import panars as pa

# Create a DataFrame
df = pa.DataFrame({
    "A": [1, 2, 3, 4],
    "B": [5, 6, 7, 8],
    "C": [1, 1, 2, 2]
})

# Familiar Pandas operations
print(df.head())
print(df.groupby("C").sum())
print(df.filter(df["A"] > 2))

# Efficient data manipulation
result = (df.groupby(["C"])
            .agg({"A": ["mean", "sum"], "B": ["min", "max"]})
            .sort_values("C"))
print(result)

Why panars?

1. Familiar Syntax: Write Polars code using Pandas conventions you already know.
2. Performance Boost: Gain Polars' speed advantages without learning a new API.
3. Gradual Migration: Easily port existing Pandas projects to Polars over time.
4. Community-Driven: Open-source project welcoming contributions and feedback.

Documentation

For detailed usage instructions and API reference, visit our documentation.

ToDo some api

dataframe

IO

• [x] read_csv
• [x] to_csv
• [x] read_excel
• [x] to_excel
• [x] read_parquet
• [x] to_parquet

Data manipulations

• [x] groupby
• [x] melt
• [x] pivot
• [x] merge
• [x] concat

Top-level missing data

• [x] isnull

Top-level dealing with datetimelike dat

• [x] to_datetime
• [x] agg
• [x] map
• [x] apply
• [x] rename

info

• [x] head
• [x] tail
• [x] info
• [x] describe

groupby

• [x] sum
• [x] mean
• [x] max
• [x] min
• [x] count

other

• [x] fillna
• [x] dropna
• [x] drop
• [x] len

Contributing

We welcome contributions! Please see our Contributing Guide for more details.

License

Panars is released under the MIT License. See the LICENSE file for details.