pandasnet is a python package build on top of pythonnet. It provides additional data conversions for pandas, numpy and datetime

• python 3.6 or higher.
• dotnet.

dotnet also provides scripts to proceed the installation by command line.

pip install pandasnet

To load the converter you need to import the package once in your python environment. If the dotnet clr isn't started yet through the pytonnet package the import will.

import pandasnet

We construct a simple C# function to test conversion

using System;
using System.Collections.Generic;

namespace LibForTests
{
    public class PandasNet
    {
        public static Dictionary<string, Array> BasicDataFrame(Dictionary<string, Array> df)
            => df;
    }
}

We build this function into a library named LibForTests.dll. We load this library into our python environment then use it.

import clr
import pandasnet # Load the converters
import pandas as pd
from datetime import datetime

# Load your dll
clr.AddReference('LibForTests.dll')
from LibForTests import PandasNet as pdnet

x = pd.DataFrame({
    'A': [1, 2, 3],
    'B': [1.23, 1.24, 1.22],
    'C': ['foo', 'bar', 'other'],
    'D': [datetime(2021, 1, 22), datetime(2021, 1, 23), datetime(2021, 1, 24)]
})
y = pdnet.BasicDataFrame(x)

print(y)

Below an exhausitve list of supported data convertions.

Issue tracker: https://github.com/fdieulle/pandasnet/issues

If you want to checkout the project and propose your own contribution, you will need to setup it following few steps:

python -m venv venv

venv/Scripts/activate

pip install -r requirements.txt

This project is open source under the MIT license.