How to append an empty array by multiple 1D array

Question

I want to create 2D array by multiple 1D array (1,7680) to have multiple number of arrays under each other creating 2D array (n,7680)

Any help would be appreciated

code

y=[]
t=0
movement=int(S*256)
if(S==0):
    movement=_SIZE_WINDOW       
while data.shape[1]-(t*movement+_SIZE_WINDOW) > 0:
    for i in range(0, 22):
        start = t*movement
        stop = start+_SIZE_WINDOW
        signals[i,:]=data[i,start:stop]
        y=np.append(signals[i,:],y)

     t=t+1

You wrote np.empty(()). Are you using numpy? If that is true maybe you would need to add the numpy tag to your question. — Ender Look
– Ender Look, Commented Jun 21, 2020 at 21:37
@AnnZen I want to create 2D array by multiple 1D array (1,7680) to have multiple number of arrays under each other creating 2D array — Eda
– Eda, Commented Jun 21, 2020 at 21:40
Appending arrays to other arrays one by one is an extremely inefficient way to use NumPy, due to the quadratic amount of copying involved. Build up a list of arrays one by one and then convert to a single array at the end, or preallocate the whole array up front and fill in chunks one by one. — user2357112
– user2357112, Commented Jun 21, 2020 at 21:41
@user2357112supportsMonica How can I do that ?? I want to create 2D array by multiple 1D array to have multiple number of arrays under each other creating 2D array — Eda
– Eda, Commented Jun 21, 2020 at 21:44

Edo98 · Accepted Answer · 2020-06-21 22:52:19Z

1

If the shape of the arrays you want to create is well defined the easiest and optimal way to do so is to create an empty array like this:

array_NxM = np.empty((N,M))

This will create an empty array with the desired shape, then you can fill the array by iterating through its elements.

Creating an array by appending 1d arrays is definitely not optimal but an acceptable way to do so would be to create a list, appending 1d arrays to it and then cast the list to a numpy array like this:

array_NxM = []
for i in range(M):
    array_NxM.append(array_1xM)
array_NxM = np.array(array_NxM)

The worst way to do this is definitely to use np.append. If possible always avoid appending to a numpy array as this operations leads to a full copy in memory of the array and a read/write of it.

answered Jun 21, 2020 at 22:52

Edo98

4532 silver badges9 bronze badges

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

4 Comments

Eda Over a year ago

I add this to my code

array_NxM = []             for J in range(7681):                 array_NxM.append(data[i,start:stop])                 array_NxM = np.array(array_NxM)

but I got this error array_NxM.append(data[i,start:stop]) AttributeError: 'numpy.ndarray' object has no attribute 'append'

Edo98 Over a year ago

You cannot use append to a numpy array. Just use the first solution, it will produce what you need

Eda Over a year ago

I don't know the shape of the array to use this array_NxM = np.empty((N,M)) .

Eda Over a year ago

What is the first solution ??

Collectives™ on Stack Overflow

How to append an empty array by multiple 1D array

1 Answer 1

4 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Related