larray.diag

larray.diag(a, k=0, axes=(0, 1), ndim=2, split=True) → Array[source]

Extract a diagonal or construct a diagonal array.

Parameters

aArray: If a has 2 dimensions or more, return a copy of its k-th diagonal. If a has 1 dimension, return an array with ndim dimensions on the k-th diagonal.
kint, optional: Offset of the diagonal from the main diagonal. Can be positive or negative. Defaults to main diagonal (0).
axestuple or list or AxisCollection of axes references, optional: Axes along which the diagonals should be taken. Use None for all axes. Defaults to the first two axes (0, 1).
ndimint, optional: Target number of dimensions when constructing a diagonal array from an array without axes names/labels. Defaults to 2.
splitbool, optional: Whether to try to split the axis name and labels. Defaults to True.

Returns

Array: The extracted diagonal or constructed diagonal array.

Examples

>>> nat = Axis('nat=BE,FO')
>>> sex = Axis('sex=M,F')
>>> a = ndtest([nat, sex], start=1)
>>> a
nat\sex  M  F
     BE  1  2
     FO  3  4
>>> d = diag(a)
>>> d
nat_sex  BE_M  FO_F
            1     4
>>> diag(d)
nat\sex  M  F
     BE  1  0
     FO  0  4
>>> a = ndtest(sex, start=1)
>>> a
sex  M  F
     1  2
>>> diag(a)
sex\sex  M  F
      M  1  0
      F  0  2