Solve sparse matrix efficiently, Python, SciPy 0.12

Question

In another message about resizing a sparse matrix in SciPy, the accepted answer works when you need to add more rows or columns using scipy.sparse.vstack or hstack , respectively. In SciPy 0.12, reshape or set_shape methods are still not implemented.

Are there any stable good practices for modifying a sparse matrix in SciPy 0.12? It would be nice to make some time comparisons.

Answer 1

I don't know any established good practices, so here's a pretty straight forward change function for coo_matrix. It converts its argument to coo_matrix, so it will work in other sparse formats (but it returns coo_matrix).

 from scipy.sparse import coo_matrix def reshape(a, shape): """Reshape the sparse matrix `a`. Returns a coo_matrix with shape `shape`. """ if not hasattr(shape, '__len__') or len(shape) != 2: raise ValueError('`shape` must be a sequence of two integers') c = a.tocoo() nrows, ncols = c.shape size = nrows * ncols new_size = shape[0] * shape[1] if new_size != size: raise ValueError('total size of new array must be unchanged') flat_indices = ncols * c.row + c.col new_row, new_col = divmod(flat_indices, shape[1]) b = coo_matrix((c.data, (new_row, new_col)), shape=shape) return b 

Example:

 In [43]: a = coo_matrix([[0,10,0,0],[0,0,0,0],[0,20,30,40]]) In [44]: aA Out[44]: array([[ 0, 10, 0, 0], [ 0, 0, 0, 0], [ 0, 20, 30, 40]]) In [45]: b = reshape(a, (2,6)) In [46]: bA Out[46]: array([[ 0, 10, 0, 0, 0, 0], [ 0, 0, 0, 20, 30, 40]]) 

Now I'm sure that there are several regular participants who can come up with something better (faster, more memory, less filling ... :)

Answer 2

I have one working example for a CSR matrix, but I can not guarantee that it always works

flattening matrix A:

  indices = zeros_like(A.indices) indices[A.indptr[1:-1]] = A.shape[1] indices = cumsum( indices)+A.indices A_flat = sparse.csc_matrix((T_rot.data, indices,[0,size(A)]),shape=(prod(A.shape),1)) 

reformatting matrix A

  indices = zeros_like(A.indices) indices[A.indptr[1:-1]] = A.shape[1] indices = cumsum( indices)+A.indices indices %= N*A.shape[1] indptr = r_[0, where(diff(indices)<0)[0]+1, size(A)] A_reshaped = sparse.csc_matrix((A.data, indices,indptr),shape=(N*A.shape[1],A.shape[0]/N )) 

Answer 3

According to SciPy 1.1.0 , the reshape and set_shape methods were implemented for all types of sparse matrices. Signatures are what you expect and they are identical to the equivalent methods in NumPy as much as possible (for example, you cannot change the shape of a vector or tensor).

Signature:

 reshape(self, shape: Tuple[int, int], order: 'C'|'F' = 'C', copy: bool = False) -> spmatrix 

Example:

 >>> from scipy.sparse import csr_matrix >>> A = csr_matrix([[0,0,2,0], [0,1,0,3]]) >>> print(A) (0, 2) 2 (1, 1) 1 (1, 3) 3 >>> B = A.reshape((4,2)) >>> print(B) (1, 0) 2 (2, 1) 1 (3, 1) 3 >>> C = A.reshape((4,2), order='F') >>> print(C) (0, 1) 2 (3, 0) 1 (3, 1) 3 

Full disclosure: I wrote implementations.