3.6. Appending elementsΒΆ

The size of a Tensor can be expanded using Tensor.append.

One can append a scalar to a rank-1 Tensor. For example:

  • In Python:

1A = cytnx.ones(4)
2print(A)
3A.append(4)
4print(A)

  • In C++:

1auto A = cytnx::ones(4);
2cout << A << endl;
3A.append(4);
4cout << A << endl;

Output >>

Total elem: 4
type  : Double (Float64)
cytnx device: CPU
Shape : (4)
[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]




Total elem: 5
type  : Double (Float64)
cytnx device: CPU
Shape : (5)
[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 4.00000e+00 ]

Note

It is not possible to append a scalar to a Tensor with rank > 1, as this operation is by itself ambiguous.

For Tensors with rank > 1, you can append a Tensor to it, provided the shape is matching. This operation is equivalent to numpy.vstack.

For example, consider a Tensor with shape (3,4,5). You can append a Tensor with shape (4,5) to it, and the resulting output will have shape (4,4,5).

  • In Python:

1A = cytnx.ones([3,4,5])
2B = cytnx.ones([4,5])*2
3print(A)
4print(B)
5
6A.append(B)
7print(A)

  • In C++:

1auto A = cytnx::ones({3, 4, 5});
2auto B = cytnx::ones({4, 5}) * 2;
3cout << A << endl;
4cout << B << endl;
5
6A.append(B);
7cout << A << endl;

Output >>

Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]
 [[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]
 [[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]]




Total elem: 20
type  : Double (Float64)
cytnx device: CPU
Shape : (4,5)
[[2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
 [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
 [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
 [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]]




Total elem: 80
type  : Double (Float64)
cytnx device: CPU
Shape : (4,4,5)
[[[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]
 [[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]
 [[1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]
  [1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 1.00000e+00 ]]
 [[2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
  [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
  [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]
  [2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 2.00000e+00 ]]]

Note

  1. The Tensor to be appended must have the same shape as the Tensor to append to, but with one index (the first one) less.

  2. You cannot append a complex type scalar/Tensor to a real type Tensor.