Python array shape difference

Question

I have an array as :

A
Out[159]: 
array([[ 1,  2,  3,  4,  5,  6,  7,  8],
       [ 2,  4,  6,  8, 10, 12, 14, 16]])

Now if I find the shape of :

A[0:1].shape
Out[150]: (1, 8)

While, if i try the same for

A[ [0,0,1,1], [0,3,2,5] ].shape
Out[151]: (4,)

First array seems to be 1 row and 8 columns and the second one seems to be 1 row and 4 columns, then why is the second answer shown as (4,) and not as (1,4) ?

shx2 · Accepted Answer · 2017-02-27 08:52:50Z

1

The second case doesn't mean what you think it means.

The way this kind of fancy-indexing works in numpy is as follows: If A is a 2dim array and I1 = [a1, a2, ...] and I2 = [b1, b2, ...] are arrays/lists of ints, this indexing:

A[I1,I2]

means:

np.array([ A[a1,b1], A[a2,b2], A[a3,b3], ... ])

I1 refers to indexes along dim=0, and I2 refers to the corresponding indexes along dim=1. This means each [a_k, b_k] pair defines a single element to include in sliced array.

Another way to think about this: think about the tuple returned by np.where, and what you expect this to mean:

A[np.where(A == x)]

answered Feb 27, 2017 at 8:52

shx2

64.8k17 gold badges139 silver badges166 bronze badges

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

2 Comments

Sarang Manjrekar Over a year ago

but then what does Out[151]: (4,) mean actually ?

shx2 Over a year ago

@Sarang it means a 1dim array of size 4. Since in your case len(I1)==len(I2)==4, you "pick" 4 elements from A

Collectives™ on Stack Overflow

Python array shape difference

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related