Converting one-dimensional structured array to 2D numpy array

Question

I have a one-dimensional structured array e.g.([ (1,2), (3,4) ]) and want to convert it to 2D numpy array ([ [1,2], [3,4] ]). Right now I am using list comprehension and then np.asarray()

list_of_lists = [list(elem) for elem in array_of_tuples]
array2D = np.asarray(list_of_lists)

This doesn't look very efficient - is there a better way? Thanks.

NOTE: the initial version of this question was mentioning "numpy array of tuples" instead of one-dimensional structured array. This might be useful for confused python newbies as me.

Why not directly use np.asarray on array_of_tuples, won't that work? — Divakar
– Divakar, Commented Sep 24, 2015 at 16:41
... How did you get an array of 2-tuples in the first place? :-) — mgilson
– mgilson, Commented Sep 24, 2015 at 16:42
Is your "numpy array of tuples" a one-dimensional structured array, or a numpy array with dtype object, or something else? — Warren Weckesser
– Warren Weckesser, Commented Sep 24, 2015 at 16:45
@Divakar np.asarray gives the same output as input, cause the input is also a numpy array — shiftyscales
– shiftyscales, Commented Sep 24, 2015 at 16:48

Warren Weckesser · Accepted Answer · 2015-09-24 17:01:42Z

3

In a comment, you stated that the array is one-dimensional, with dtype [('x', '<f4'), ('y', '<f4'), ('z', '<f4'), ('confidence', '<f4'), ('intensity', '<f4')]. All the fields in the structured array are the same type ('<f4'), so you can create a 2-d view using the view method:

In [13]: x
Out[13]: 
array([(1.0, 2.0, 3.0, 4.0, 5.0), (6.0, 7.0, 8.0, 9.0, 10.0)], 
      dtype=[('x', '<f4'), ('y', '<f4'), ('z', '<f4'), ('confidence', '<f4'), ('intensity', '<f4')])

In [14]: x.view('<f4').reshape(len(x), -1)
Out[14]: 
array([[  1.,   2.,   3.,   4.,   5.],
       [  6.,   7.,   8.,   9.,  10.]], dtype=float32)

In this case, the view method gives a flattened result, so we have to reshape it into a 2-d array using the reshape method.

answered Sep 24, 2015 at 17:01

Warren Weckesser

116k20 gold badges207 silver badges224 bronze badges

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

1 Comment

shiftyscales Over a year ago

Thanks this works like a charm. I shall edit my question above to state that it is a structured array.

Collectives™ on Stack Overflow

Converting one-dimensional structured array to 2D numpy array

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related