diags_array#
- scipy.sparse.diags_array(diagonals, /, *, offsets=0, shape=None, format=None, dtype=<object object>)[source]#
Construct a sparse array from diagonals.
- Parameters:
- diagonalssequence of array_like
Sequence of arrays containing the array diagonals, corresponding to offsets.
- offsetssequence of int or an int, optional
- Diagonals to set (repeated offsets are not allowed):
k = 0 the main diagonal (default)
k > 0 the kth upper diagonal
k < 0 the kth lower diagonal
- shapetuple of int, optional
Shape of the result. If omitted, a square array 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 array format is returned. This choice is subject to change.
- dtypedtype, optional
Data type of the array. If dtype is None, the output data type is determined by the data type of the input diagonals.
Up until SciPy 1.19, the default behavior will be to return an array with an inexact (floating point) data type. In particular, integer input will be converted to double precision floating point. This behavior is deprecated, and in SciPy 1.19, the default behavior will be changed to return an array with the same data type as the input diagonals. To adopt this behavior before version 1.19, use dtype=None.
- Returns:
- new_arraydia_array
dia_array
holding the values in diagonals offset from the main diagonal as indicated in offsets.
See also
dia_array
constructor for the sparse DIAgonal format.
Notes
Repeated diagonal offsets are disallowed.
The result from
diags_array
is the sparse equivalent of:np.diag(diagonals[0], offsets[0]) + ... + np.diag(diagonals[k], offsets[k])
diags_array
differs fromdia_array
in the way it handles off-diagonals. Specifically,dia_array
assumes the data input includes padding (ignored values) at the start/end of the rows for positive/negative offset, whilediags_array
assumes the input data has no padding. Each value in the input diagonals is used.Added in version 1.11.
Examples
>>> from scipy.sparse import diags_array >>> diagonals = [[1.0, 2.0, 3.0, 4.0], [1.0, 2.0, 3.0], [1.0, 2.0]] >>> diags_array(diagonals, offsets=[0, -1, 2]).toarray() array([[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_array([1.0, -2.0, 1.0], offsets=[-1, 0, 1], shape=(4, 4)).toarray() array([[-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_array([1.0, 2.0, 3.0], offsets=1).toarray() array([[ 0., 1., 0., 0.], [ 0., 0., 2., 0.], [ 0., 0., 0., 3.], [ 0., 0., 0., 0.]])