通过匹配多行排序 Pandas Dataframe

Question

Suppose I have a DataFrame like this:

data=np.array([[-1.5625e-05,-1.5625e-05,-4.6875e-05],
[-1.5625e-05,-1.5625e-05,-1.5625e-05],
[-1.5625e-05,1.5625e-05,-4.6875e-05],
[-1.5625e-05,1.5625e-05,-1.5625e-05],
[1.5625e-05,-1.5625e-05,-4.6875e-05],
[1.5625e-05,-1.5625e-05,-1.5625e-05],
[1.5625e-05,1.5625e-05,-4.6875e-05],
[1.5625e-05,1.5625e-05,-1.5625e-05]])

df=pd.DataFrame(data=data,columns=['x','y','z'])

and a numpy array

coord=np.array([[-1.5625e-05,-1.5625e-05,-4.6875e-05],
[-1.5625e-05,1.5625e-05,-4.6875e-05],
[1.5625e-05,-1.5625e-05,-4.6875e-05],
[1.5625e-05,1.5625e-05,-4.6875e-05],
[-1.5625e-05,-1.5625e-05,-1.5625e-05],
[-1.5625e-05,1.5625e-05,-1.5625e-05],
[1.5625e-05,-1.5625e-05,-1.5625e-05],
[1.5625e-05,1.5625e-05,-1.5625e-05]])

The number of rows in the Pandas Dataframe and the coord array are always the same. As you can see, the rows between the Pandas DataFrame and the coord are the same but in a different order. I would like to sort the DataFrame according to the order of the coord array(eg df.x==coord[:,0] & df.y==coord[:,1] & df.z==coord[:,2]).

Answer 1

You can do the required sorting by:

df.sort_values(['x', 'y', 'z'], ascending=[True, True, True])

The full code:

import numpy as np
import pandas as pd

data=np.array([[-1.5625e-05,-1.5625e-05,-4.6875e-05],
[-1.5625e-05,-1.5625e-05,-1.5625e-05],
[-1.5625e-05,1.5625e-05,-4.6875e-05],
[-1.5625e-05,1.5625e-05,-1.5625e-05],
[1.5625e-05,-1.5625e-05,-4.6875e-05],
[1.5625e-05,-1.5625e-05,-1.5625e-05],
[1.5625e-05,1.5625e-05,-4.6875e-05],
[1.5625e-05,1.5625e-05,-1.5625e-05]])

df= pd.DataFrame(data=data,columns=['x','y','z'])

df.sort_values(['x', 'y', 'z'], ascending=[True, True, True])

print(df)

the output:

          x         y         z
0 -0.000016 -0.000016 -0.000047
1 -0.000016 -0.000016 -0.000016
2 -0.000016  0.000016 -0.000047
3 -0.000016  0.000016 -0.000016
4  0.000016 -0.000016 -0.000047
5  0.000016 -0.000016 -0.000016
6  0.000016  0.000016 -0.000047
7  0.000016  0.000016 -0.000016

Answer 2

you could do it like this:

• sort both arrays the same way
• set the index of the dataframe to that of coord
• reset the index to get the original ordering:

code

df2 = pd.DataFrame(coord, columns=list("xyz"))
sort_cols = list("yxz")
df = df.sort_values(sort_cols)
df2 = df2.sort_values(sort_cols)
df.index = df2.index
df = df.sort_index()

This should returns the df sorted as coord (output):

          x         y         z
0 -0.000016 -0.000016 -0.000047
1 -0.000016  0.000016 -0.000047
2  0.000016 -0.000016 -0.000047
3  0.000016  0.000016 -0.000047
4 -0.000016 -0.000016 -0.000016
5 -0.000016  0.000016 -0.000016
6  0.000016 -0.000016 -0.000016
7  0.000016  0.000016 -0.000016