dplyr R 在 Pandas 中排列等效函数

Question

I have a data frame similar to this, my idea is to arrange the rows according to the vector my_order , as shown below.

R Code:

df = data.frame(A = c("apple","cherry","orange","banana"), B = c(25,37,15,28))
df
       A  B
1  apple 25
2 cherry 37
3 orange 15
4 banana 28

my_order = c(2,3,4,1)
dplyr::arrange(df,my_order)
       A  B
1 banana 28
2  apple 25
3 cherry 37
4 orange 15

My question is how can I do this in python, is there any function in pandas, equivalent dplyr::arrange() ?

Python Code:

import pandas as pd

df = pd.DataFrame({'A': ["apple","cherry","orange","banana"], 'B': [25,37,15,28]})
print(df)
        A   B
0   apple  25
1  cherry  37
2  orange  15
3  banana  28

my_order = [1,2,3,0]
df.iloc[my_order]
        A   B
1  cherry  37
2  orange  15
3  banana  28
0   apple  25

Answer 1

Okay, I figured it out. You are passing argsorted indices to arrange . You can do the same thing with iloc , but you will have to argsort your indices to get its inverse.

my_order = [2,3,4,1]
df.iloc[pd.np.argsort(my_order)]

        A   B
3  banana  28
0   apple  25
1  cherry  37
2  orange  15

Answer 2

I am not sure about the right function.

work around:

import pandas as pd

df = pd.DataFrame({'A': ["apple","cherry","orange","banana"], 'B': [25,37,15,28]})

print(df)

df['index']=[2,3,4,1]
df.set_index('index',inplace=True)
df.sort_index(inplace=True)

print(df)

Answer 3

Check with

df.loc[pd.Series(my_order,index=df.index).sort_values().index]
Out[42]: 
        A   B
3  banana  28
0   apple  25
1  cherry  37
2  orange  15

Answer 4

Now you don't have to learn APIs of pandas to turn your R code into python code!

With datar :

>>> from datar import f
>>> from datar.tibble import tibble
>>> from datar.dplyr import arrange
>>> df = tibble(A = ["apple","cherry","orange","banana"], B = [25,37,15,28])
>>> df
        A   B
0   apple  25
1  cherry  37
2  orange  15
3  banana  28
>>> my_order = [2,3,4,1]
>>> df >> arrange(my_order)
        A   B
0  banana  28
1   apple  25
2  cherry  37
3  orange  15

I am the author of the package. Feel free to submit issues if you have any questions.