README.md

datar

A Grammar of Data Manipulation in python

Documentation | Reference Maps | Notebook Examples | API

datar is a re-imagining of APIs for data manipulation in python with multiple backends supported. Those APIs are aligned with tidyverse packages in R as much as possible.

Installation

pip install -U datar

# install with a backend
pip install -U datar[pandas]

# More backends support coming soon

Backends

Repo	Badges
datar-numpy
datar-pandas
datar-arrow

Example usage

# with pandas backend
from datar import f
from datar.dplyr import mutate, filter_, if_else
from datar.tibble import tibble
# or
# from datar.all import f, mutate, filter_, if_else, tibble

df = tibble(
    x=range(4),  # or c[:4]  (from datar.base import c)
    y=['zero', 'one', 'two', 'three']
)
df >> mutate(z=f.x)
"""# output
        x        y       z
  <int64> <object> <int64>
0       0     zero       0
1       1      one       1
2       2      two       2
3       3    three       3
"""

df >> mutate(z=if_else(f.x>1, 1, 0))
"""# output:
        x        y       z
  <int64> <object> <int64>
0       0     zero       0
1       1      one       0
2       2      two       1
3       3    three       1
"""

df >> filter_(f.x>1)
"""# output:
        x        y
  <int64> <object>
0       2      two
1       3    three
"""

df >> mutate(z=if_else(f.x>1, 1, 0)) >> filter_(f.z==1)
"""# output:
        x        y       z
  <int64> <object> <int64>
0       2      two       1
1       3    three       1
"""

# works with plotnine
# example grabbed from https://github.com/has2k1/plydata
import numpy
from datar import f
from datar.base import sin, pi
from datar.tibble import tibble
from datar.dplyr import mutate, if_else
from plotnine import ggplot, aes, geom_line, theme_classic

df = tibble(x=numpy.linspace(0, 2 * pi, 500))
(
    df
    >> mutate(y=sin(f.x), sign=if_else(f.y >= 0, "positive", "negative"))
    >> ggplot(aes(x="x", y="y"))
    + theme_classic()
    + geom_line(aes(color="sign"), size=1.2)
)

# very easy to integrate with other libraries
# for example: klib
import klib
from pipda import register_verb
from datar import f
from datar.data import iris
from datar.dplyr import pull

dist_plot = register_verb(func=klib.dist_plot)
iris >> pull(f.Sepal_Length) >> dist_plot()

Testimonials

@coforfe:

Thanks for your excellent package to port R (dplyr) flow of processing to Python. I have been using other alternatives, and yours is the one that offers the most extensive and equivalent to what is possible now with dplyr.