在数组中找到与另一个数组python中的值相对应的行

Question

I have the following arrays:

arr1 = np.array(['x','y','z'])

arr2 = np.array(['y','a','b','z','c','x'])

arr3 = np.array([['y','d1','d2','d3'],
                 ['a','d4','d5','d6'],
                 ['b','d7','d8','d9'],
                 ['z','d10','d11','d12'],
                 ['c','d13','d14','d15'],
                 ['x','d16','d17','d18']])

and I want to get the following:

answer = np.array([[x,d16,d17,d18],
                   [y,d1,d2,d3]
                   [z,d10,d11,d12]])

How do I do this??

Answer 1

I think this should work, or at least get you in the right direction.

idxs = np.where( arr3[:,0] == arr1 )
answer = arr3[idxs, :]

read about np.where: http://docs.scipy.org/doc/numpy/reference/generated/numpy.where.html

Answer 2

inds = [ where( arr3[:,0] == val)[0] for val in arr1 ]
inds = [ i[0] for i in inds if i.size ] 
answer = arr3[inds]

Answer 3

Take the first row ( next() ), where the first element equals an element e1 of arr1 .

array([
    next(row3 for row3 in arr3 if row3[0]==e1)
      for e1 in arr1])

Out:

array([['x', 'd16', 'd17', 'd18'],
       ['y', 'd1', 'd2', 'd3'],
       ['z', 'd10', 'd11', 'd12']], 
  dtype='|S3')

Answer 4

The numpy_indexed package (disclaimer: I am its author) has functionality to address this problem efficiently. You can think of it as a vectorized equivalent of list.index. Should be in numpy itself if you ask me; maybe in a future version!

import numpy_indexed as npi
answer = arr3[npi.indices(arr3[:,0], arr1)]