scisharp / pandas.net

Pandas port for C# and F#, data analysis tool, process multi-dim array in DataFrame.

Pandas.NET

Implemented APIs

1. Pandas

• DataFrame
  • pd.DataFrame(NDArray data, IList<string> index, IList<string> columns, Type dtype)
  • pd.DataFrame<TIndex>(NDArray data, IList<TIndex> index, IList<string> columns, Type dtype)
  • pd.DataFrame(IDictionary<string, NDArray> data, IList<string> index)
  • pd.DataFrame<TIndex>(IDictionary<string, NDArray> data, IList<TIndex> index)
• Series
  • pd.Series(NDArray data)
  • pd.Series(Array data)
  • pd.Series<T>(T data)

2. Series

• s.iloc[0]: Select data by index

• s.loc["index_label"]： Select data by index label

3. DataFrame

Structure

• df.Index
• df.Columns
• df.Values
• df.Shape
• df.NDIM
• df.Size

Method

• df[0]： Select data by column index （returns Series）

• df[params int[] columnIndexs] : Select data by column index （returns DataFrame）

• df["column_label"]： Select data by column label （returns Series）; accessible set accessor increase column （if the column label does not exist）

• df[params string[] columnLabels]： Select data by column label （returns DataFrame）

• df.Column(string columnLabel, NDArray value): Set the column and its value; create when the column does not exist

• df.Column(int columnIndex, NDArray value): Set the column and the value of the column; when the column does not exist, an exception is reported

• df[Slice s]: Row slice selection data

• df.loc["index_label"]： Select data by row index label

• df.loc["index_label", "column_label"]： Select data by row and column labels

• df.iloc[0]： Index by row （row number） select data

Libraries

• Math.NET Numerics
  • MIT License

pandas.net's Issues

iloc and loc from Pandas

Hello there is a compatible function with Pandas iLoc and Loc ?

Rationale - The Second Deedle

Please provide rationale somewhere in main README, why using/developing/participating in this project rather than Deedle.

Can you put a DLL on nuget?

在nuget上找不到这个库的dll,而自己下载下来编译后测试，好多报错，不知道是不是自己操作不当的原因。
水平比较差，不懂太多，只想有个.net版本的可以用一下，我主要是做Excel开发的，想做点自定义函数出来。

Open a SciSharp repo

With SciSharp we would have at least all 3 repos started from scipy stack. Could you open a repo for this?

Honesty I am very sure that SciSharp is also quite important. Regression, signal processing etc.

I don't want to start it in my account since it would be confusing if 2 repos at your side and 1 is mine lol

Worth a look - Deedle

I remembered this library from my F# days.
http://bluemountaincapital.github.io/Deedle/index.html

It does some things pandas does and it's dotnet. It specifically handles DataFrames.

This link is more helpful to C# users
http://bluemountaincapital.github.io/Deedle/csharpframe.html

Implement pandas.read_csv

Read CSV (comma-separated) file into DataFrame.
https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_csv.html

Where is "iloc"?

This repository has README.MD and that describes about "iloc" function.
But I can not find.
Could you let me know how I can use that?

致命Bug着实有点多，都运行不起来

有活干了，抽空我要好好修修bug了

Create Series From ndarray

If data is an ndarray, index must be the same length as data. If no index is passed, one will be created having values [0, ..., len(data) - 1].

In:

s = pd.Series(np.random.randn(5), index=['a', 'b', 'c', 'd', 'e'])

Out:

a    0.4691
b   -0.2829
c   -1.5091
d   -1.1356
e    1.2121
dtype: float64

https://pandas.pydata.org/pandas-docs/stable/dsintro.html

Question : Should we start Pandas.NET in NumSharp repo?

It is totally great to start Pandas.NET.

But maybe we should discuss to make it as a project in NumSharp.

So the structure (of big picture in future) would be like that :

Repo NumSharp (maybe we can call it later as "NumSharpStack" so it is clear all libraries that bases on NumSharp).

/src/NumSharp/Numsharp.csproj
/src/Pandas.NET/Pandas.NET.csproj
/src/SciSharp/SciSharp.csproj

The reason for that is --> at moment we are still under construction.
If NumSharp change we do not know what impact it has for Pandas.NET. Moreover
at moment we could build all stuff with command "dotnet build" and we do not depend on
Nuget package.

But thats up to u. :)

All unit tests failed

I just saw this great library, but seems the current implementation is not stable. I run unit tests but all tests failed.
What is the current status of the library?

Implement Series.sample

Return a random sample of items from an axis of object.
https://pandas.pydata.org/pandas-docs/stable/generated/pandas.Series.sample.html

Provide the license for using this library

As we have License specified for other SciSharp repositories, we need to provide the License section and declare its type in our repository readme file.
Provide the opportunity for the 3rd party or users to make correct integration with our product without any legal obligations.

Test suite is lacking many tests

PandasNet needs a suite of tests built on a modern test platform for all functions including proper validation with Asserts.

Readme in English

Can someone please convert the Chinese text in the readme to English.

Please note: the major language for all SciSharp projects is English. Translations of the documentation in other languages should only be second to English and in a separate subfolder.

DataFrame构造函数初始化报错

用的是单元测试里的代码，但报错。

CSV Reader + Span<T>

Good read here with repo here. Using Span<T> will have extreme performance and very few extra allocations.

.net 6.0 EOL Nov 8, 2024

Upgrade packages to .NET 8

Non Pandas function arange

Does the arange function belong in the Pandas.NET framework? I believe this is strictly a NumSharp function. Would it make sense to simply offer a forwarded call to NumSharp from PandasNET vs implementing it's own?

Melt/Pivot function

Hi, Im testing the package to be used in a new application.
I couldn't find MELT()/PIVOT() function from pandas.
Did you implemented it on Pandas.NET?

Run on Multi-Core CPUs, GPUs, or a cluster

Cannot able to open in VS 2015

Should DataFrame really be a NDArray child?

Sorry I already rise an issue while all is under construction >.<.

We should not let dataframe be a child of ndarray.
In Pandas the dataframe is a child of a general pandas object and has no inheritance connection to NDArray.
I think we will face same problems if we in heritage from ndarray.

• Data frame is a collection of multiple NDArrays so it is not a array itself it is a collection
• the behaviour is different from an array especially in indexing
• frame indexing is like a dictionary df['column1'] or df['col2']
• honestly spoken I think we need a dictionary object as property so we can store the different arrays
• the indexing of dataframe must be via strings like in dictionary (not sure if possible but i think it must be possible)

Implement Series.head

Return the first n rows.
https://pandas.pydata.org/pandas-docs/stable/generated/pandas.Series.head.html