How to sort a numpy array based on the both arrays and also conditions

Question

I have a numpy array with three columns and hundres of rows. I want to sort it based on two columns and in cases also regard the third column in my sorting. This is my input (they are x, y and z coordinates):

my_point=np.array([[1., 2., 90.9],
                   [1., 0., 100.1],
                   [1.8, 0., 2.8],
                   [1.8, 2., 3.1],
                   [1.8, 1., 2.7],
                   [3., 0., 3.],
                   [3., 2., 2.9],
                   [2., 1., 100.],
                   [2., 0., 100.]])

Firstly, I want to sort it based on the first (x) and second column (y) and secondly modify this sort based on the third column (z). The key issue is that I must sort my data based on x and y and cannot change it to x and z or y and z, because then it changes the order of all the points. This is my sorting code:

result_array=my_point[np.lexsort((my_point[:,1],my_point[:,0]))]

It gives me:

array([[1., 0., 100.1],
       [1., 2., 90.9],
       [1.8, 0., 2.8],
       [1.8, 1., 2.7],
       [1.8, 2., 3.1],
       [2., 0., 100.],
       [2., 1., 100.],
       [3., 0., 3. ],
       [3., 2., 2.9]])

It is correct but I want to modify my sort based on the third column. I want to consider the values of this column, when it is highly different from the adjacent rows. As result_array shows, third column of its second, third and fourth rows ([1.8, 0., 2.8], [1.8, 1., 2.7] and [1.8, 2., 3.1]) have very low values compared to the next two rows. The x value of these rows is just a little bit higher than next ones and in such cases I want to neglect it and give priority to z values:

array([[1., 0., 100.1],
       [1., 2., 90.9],
       [2., 0., 100.],
       [2., 1., 100.],
       [1.8, 0., 2.8],
       [1.8, 1., 2.7],
       [1.8, 2., 3.1],
       [3., 0., 3. ],
       [3., 2., 2.9]])

For more clarity, I have uploaded a fig showing my desired order of points in 3d space. I think It can be done maybe through defining a function for sorting but I have no idea on how to do it. In advance, I do appreciate any help.

Nils Werner · Accepted Answer · 2020-11-25 14:22:32Z

1

You could define a sort function like

sortfunc = my_point[:, 0] * 100 + my_point[:, 1] * 10 - my_point[:, 2]

idx = np.argsort(sortfunc)
my_point[idx]
# array([[  1. ,   0. , 100.1],
#        [  1. ,   2. ,  90.9],
#        [  2. ,   0. , 100. ],
#        [  2. ,   1. , 100. ],
#        [  1.8,   0. ,   2.8],
#        [  1.8,   1. ,   2.7],
#        [  1.8,   2. ,   3.1],
#        [  3. ,   0. ,   3. ],
#        [  3. ,   2. ,   2.9]])

Of course the coefficients 100, 10, and -1 are then a matter of tuning.

answered Nov 25, 2020 at 14:22

Nils Werner

37.2k7 gold badges85 silver badges108 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Link_tester Over a year ago

Dear @Nils Werner, Sorry for my late feed back. Since yesterday I was working with my data. Your method as completely fine but with the new data, I have a big problem and cannot tune the coeficients at all. Do you know any other way to do it.

Nils Werner Over a year ago

No, but I suggest you just open another question with the new data and problems.

Link_tester Over a year ago

Dear @Nils Werner, Thanks for your help. I do appreciate it.

Collectives™ on Stack Overflow

How to sort a numpy array based on the both arrays and also conditions

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related