How to Slice a 2D Array into Smaller 2D Subarrays in NumPy?

Slicing 2D Arrays into Smaller 2D Subarrays

Question:
Can we subdivide a 2D array into smaller 2D arrays in NumPy?

Example:
Transform a 2x4 array into two 2x2 arrays:

[[1,2,3,4]   ->    [[1,2] [3,4]
 [5,6,7,8]]          [5,6] [7,8]]

Mechanism:

Instead of creating new arrays, a better approach is to reshape the existing array using reshape() and swap axes using swapaxes().

Blockshaped Function:

Below is the implementation of the blockshaped function:

def blockshaped(arr, nrows, ncols):
    """
    Partitions an array into blocks.

    Args:
        arr (ndarray): The original array.
        nrows (int): Number of rows in each block.
        ncols (int): Number of columns in each block.

    Returns:
        ndarray: Partitioned array.
    """
    h, w = arr.shape
    assert h % nrows == 0, f"{h} rows is not evenly divisible by {nrows}"
    assert w % ncols == 0, f"{w} cols is not evenly divisible by {ncols}"
    return (arr.reshape(h // nrows, nrows, -1, ncols)
               .swapaxes(1, 2)
               .reshape(-1, nrows, ncols))

Demo:

np.random.seed(365)
c = np.arange(24).reshape((4, 6))
print(c)

print(blockshaped(c, 2, 3))

Alternative Solution:

SuperBatFish's blockwise_view provides another option, offering a different block arrangement and view-based representation.

The above is the detailed content of How to Slice a 2D Array into Smaller 2D Subarrays in NumPy?. For more information, please follow other related articles on the PHP Chinese website!