I think the nzrange function (new to 0.4 function) is best to do
that:
Base.nzrange(A, col)
Return the range of indices to the structural nonzero values of a
sparse matrix column. In conjunction with "nonzeros(A)" and
"rowvals(A)", this allows for convenient iterating over a sparse
matrix
A = sparse(I,J,V)
rows = rowvals(A)
vals = nonzeros(A)
m, n = size(A)
for i = 1:n
for j in nzrange(A, i)
row = rows[j]
val = vals[j]
# perform sparse wizardry...
end
end
If you're on 0.3 just add these two definitions from 0.4:
rowvals(S::SparseMatrixCSC) = S.rowval
nzrange(S::SparseMatrixCSC, col::Integer) =
S.colptr[col]:(S.colptr[col+1]-1)
(this is just what Odd suggested)
And yes, transposing is likely the fastest for doing this over rows as
indexing into columns is very expensive.
