diags_array#
- scipy.sparse.diags_array(diagonals, /, *, offsets=0, shape=None, format=None, dtype=None)[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. 
 
 - Notes - Repeated diagonal offsets are disallowed. - The result from - diags_arrayis the sparse equivalent of:- np.diag(diagonals[0], offsets[0]) + ... + np.diag(diagonals[k], offsets[k]) - diags_arraydiffers from- dia_arrayin the way it handles off-diagonals. Specifically,- dia_arrayassumes the data input includes padding (ignored values) at the start/end of the rows for positive/negative offset, while- diags_array` assumes the input data has no padding. Each value in the input ``diagonalsis used.- Added in version 1.11. - Examples - >>> from scipy.sparse import diags_array >>> diagonals = [[1, 2, 3, 4], [1, 2, 3], [1, 2]] >>> 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, -2, 1], 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, 2, 3], offsets=1).toarray() array([[ 0., 1., 0., 0.], [ 0., 0., 2., 0.], [ 0., 0., 0., 3.], [ 0., 0., 0., 0.]])