SciPy

scipy.sparse.diags

scipy.sparse.diags(diagonals, offsets, shape=None, format=None, dtype=None)[source]

Construct a sparse matrix from diagonals.

New in version 0.11.

Parameters :

diagonals : sequence of array_like

Sequence of arrays containing the matrix diagonals, corresponding to offsets.

offsets : sequence of int

Diagonals to set:
  • k = 0 the main diagonal
  • k > 0 the k-th upper diagonal
  • k < 0 the k-th lower diagonal

shape : tuple of int, optional

Shape of the result. If omitted, a square matrix large enough to contain the diagonals is returned.

format : {“dia”, “csr”, “csc”, “lil”, ...}, optional

Matrix format of the result. By default (format=None) an appropriate sparse matrix format is returned. This choice is subject to change.

dtype : dtype, optional

Data type of the matrix.

See also

spdiags
construct matrix from diagonals

Notes

This function differs from spdiags in the way it handles off-diagonals.

The result from diags is the sparse equivalent of:

np.diag(diagonals[0], offsets[0])
+ ...
+ np.diag(diagonals[k], offsets[k])

Repeated diagonal offsets are disallowed.

Examples

>>> diagonals = [[1,2,3,4], [1,2,3], [1,2]]
>>> diags(diagonals, [0, -1, 2]).todense()
matrix([[1, 0, 1, 0],
        [1, 2, 0, 2],
        [0, 2, 3, 0],
        [0, 0, 3, 4]])

Broadcasting of scalars is supported (but shape needs to be specified):

>>> diags([1, -2, 1], [-1, 0, 1], shape=(4, 4)).todense()
matrix([[-2.,  1.,  0.,  0.],
        [ 1., -2.,  1.,  0.],
        [ 0.,  1., -2.,  1.],
        [ 0.,  0.,  1., -2.]])

If only one diagonal is wanted (as in numpy.diag), the following works as well:

>>> diags([1, 2, 3], 1).todense()
matrix([[ 0.,  1.,  0.,  0.],
        [ 0.,  0.,  2.,  0.],
        [ 0.,  0.,  0.,  3.],
        [ 0.,  0.,  0.,  0.]])

Previous topic

scipy.sparse.kronsum

Next topic

scipy.sparse.spdiags