dplython

> 编程语言

开源

dplyr 适用于 Python

761 stars0 点赞0 次浏览

访问官网 GitHub

工具介绍

dplyr 适用于 Python

Dplython: Dplyr for Python

Welcome to Dplython: Dplyr for Python.

Dplyr is a library for the language R designed to make data analysis fast and easy. The philosophy of Dplyr is to constrain data manipulation to a few simple functions that correspond to the most common tasks. This maps thinking closer to the process of writing code, helping you move closer to analyze data at the "speed of thought".

The goal of this project is to implement the functionality of the R package Dplyr on top of Python's pandas.

Dplyr: Click here
Pandas: Click here

This is version 0.0.7. It's experimental and subject to change.

Introductory Video

Here is a 20 minute video explaining dplython, given at PyGotham 2016. Click the awkward picture above to see the talk! Note that sound doesn't start until about 1 minute in due to microphone issues.

Installation

To install, use pip:

bash

pip install dplython

To get the latest development version, you can clone this repo or use the command:

bash

pip install git+https://github.com/dodger487/dplython.git

Contributing

We welcome your feature requests, open issues, bug reports, and pull requests! Please use GitHub's interface. Also consider joining the dplython mailing list.

Example usage

…

python

(diamonds >>
  sift((X.clarity == "I1") | (X.clarity == "IF")) >> 
  ggplot(aes(x="carat", y="price", color="color"), X._) + 
    geom_point() + 
    facet_wrap("clarity"))

python

# Matplotlib works as well!
import pylab as pl
pl.scatter = DelayFunction(pl.scatter)
diamonds >> sample_frac(0.1) >> pl.scatter(X.carat, X.price)

This is very new and I'm matching changes. Let me know if you'd like to see a feature or think there's a better way I can do something.

Other approaches

pandas-ply

Development of dplython began before I knew pandas-ply existed. After I found it, I chose "X" as the manager to be consistent. Pandas-ply is a great approach and worth taking a look. The main contrasts between the two are that:

dplython uses dplyr-style groups, as opposed to the SQL-style groups of pandas and pandas-ply
dplython maps a little more directly onto dplyr, for example having mutate instead of an expanded select.
Use of operators to connect operations instead of method-chaining