如何将动态生成的值附加到NumPy数组中？

Question

I want to add some dynamically calculated values to result array:

out = []
for item in foo:
    spam = get_spam(foo)
    out.append(spam)
return np.array(out) 

Unfortunately, this doesn't work, because spam shape is (10,1) and out array has one extra dimension - this is bad.

I did something like this:

for index, item in enumerate(foo):
    spam = get_spam(foo)
    if index == 0:
        out = spam
    else:
        out = np.append(out, spam, 1)
return out

But this solution is not short, easy, and readable.

How do this in pythonic way?

Answer 1

You can use numpy concatenate function:

np.concatenate([get_spam(foo) for item in foo])

It eventually accepts an argument axes for the direction of concatenation (vertically/horizontally).

Answer 2

if spam always has the shape (10,1) you could slice the returned array with [:,0]

for item in foo:
    spam = get_spam(foo)
    out.append(spam[:,0])
return np.array(out)