Migration guide for NumPy.
| Description | NumPy | stdlib |
|---|---|---|
| Append a zero-filled array of the same shape along a specified dimension | np.concat((x, np.zeros_like(x)), axis=dim) |
concat([x, zerosLike(x)], {dim: dim}) |
| Broadcast an array to a specified shape | np.broadcast_to(x, shape) |
broadcastArray(x, shape) |
| Broadcast a scalar to a specified shape | np.broadcast_to(np.array(scalar), shape) |
broadcastScalar(scalar, shape) |
| Compute the maximum absolute value | np.max(np.abs(x)) |
maxabs(x) |
| Compute the minimum absolute value | np.min(np.abs(x)) |
minabs(x) |
| Concatenate a list of arrays along the second-to-last dimension | np.concat(arrays, axis=-2) |
vconcat(arrays) |
| Concatenate a list of arrays along the last dimension | np.concat(arrays, axis=-1) |
hconcat(arrays) |
| Concatenate a list of one-dimensional arrays as columns | np.column_stack(arrays) |
colcat(arrays) |
| Concatenate a list of one-dimensional arrays as rows | np.vstack(arrays) |
rowcat(arrays) |
| Copy an array | np.copy(x) |
copy(x) |
| Count the number of falsy values in an array | x.size-np.count_nonzero(x) |
countFalsy(x) |
| Count the number of truthy values in an array | np.count_nonzero(x) |
countTruthy(x) |
| Create a sorted copy of an array | np.sort(x) |
toSorted(x) |
| Create an array containing evenly spaced numbers over a specified interval and having a desired shape | np.reshape(np.linspace(start, stop), shape) |
linspace(shape, start, stop) |
| Create an array of uniformly distributed pseudorandom numbers | np.random.default_rng().uniform(low,high,shape) |
uniform(shape,low,high) |
| Filter an array according to a predicate function | x[np.vectorize(predicate)(x)] |
filter(x, predicate) |
| Find the index of the first element which equals a specified value | np.argmax(x == v, axis=dim) |
indexOf(x, v, {dim: dim}) |
| Find the index of the last element which equals a specified value | x.shape[dim]-1-np.argmax(np.flip(x, axis=dim) == v, axis=dim) |
lastIndexOf(x, v, {dim: dim}) |
| Flatten an array to a desired depth | np.reshape(x, newshape) |
flatten(x, {depth: depth}) |
| Flatten an array starting from a specific dimension | np.reshape(x, x.shape[:dim] + (-1,)) |
flattenFrom(x, dim) |
| Prepend a specified number of singleton dimensions | np.reshape(x, (1,)*n + x.shape) |
prependSingletonDimensions(x, n) |
| Prepend a zero-filled array of the same shape along a specified dimension | np.concat((np.zeros_like(x), x), axis=dim) |
concat([zerosLike(x), x], {dim: dim}) |
| Remove singleton dimensions | np.squeeze(x) |
removeSingletonDimensions(x) |
| Reverse the elements along a dimension | np.flip(x, axis=dim) |
reverseDimension(x, dim) |
| Rotate an array by 90 degrees in a specified plane | np.rot90(x, axes=dims) |
rot90(x, {dims: dims}) |
| Rotate an array by 180 degrees in a specified plane | np.rot90(x, k=2, axes=dims) |
rot180(x, {dims: dims}) |
| Sort an array in-place | x[:] = np.sort(x) |
sort(x) |
Test whether an array contains at least n truthy values |
np.count_nonzero(x) >= n |
some(x, n) |
| Test whether an array contains truthy values | np.any(x) |
any(x) |
| Test whether an array includes a specific value | np.any(x == v) |
includes(x,v) |
{{table}}