This vignette is adapted from the official Armadillo documentation.
The following table uses V for vectors, M for matrices, Q for cubes and F for fields. For operations, A, B and C denote matrices. Because MATLAB indexes from 1 (as R does), the k-th column in MATLAB has index k but in C++ it has index k-1.
| Matlab/Octave | Armadillo | Notes |
|---|---|---|
M(1, 1) |
M(0, 0) |
indexing in C++ starts at 0 |
M(k, k) |
M(k-1, k-1) |
indexing in C++ starts at 0 |
size(M,1) |
M.n_rows |
read only |
size(M,2) |
M.n_cols |
read only |
size(Q,3) |
Q.n_slices |
|
numel(A) |
M.n_elem |
|
M(:, k) |
M.col(k-1) |
|
M(k, :) |
M.row(k) |
|
M(:, p:q) |
M.cols(p, q) |
|
M(p:q, :) |
M.rows(p, q) |
|
M(p:q, r:s) |
M(span(p,q), span(r,s)) |
|
Q(:, :, k) |
Q.slice(k) |
|
Q(:, :, t:u) |
Q.slices(t, u) |
|
Q(p:q, r:s, t:u) |
Q( span(p,q), span(r,s), span(t,u)) |
|
M' |
M.t() or trans(M) |
matrix transpose / Hermitian transpose (for complex matrices, the conjugate of each element is taken) |
M = zeros(size(M)) |
M.zeros() |
|
M = ones(size(M)) |
M.ones() |
|
M = zeros(k) |
M = zeros<mat>(k,k) |
|
M = ones(k) |
M = ones<mat>(k,k) |
|
C = complex(A,B) |
cx_mat C = cx_mat(A,B) |
|
A .* B |
A % B |
element-wise multiplication |
A ./ B |
A / B |
element-wise division |
A \ B |
solve(A,B) |
more efficient than inv(A)*B |
M = M + 1 |
M++ |
|
M = M - 1 |
M-- |
|
M = [1 2; 3 4;] |
M = {{1, 2}, {3, 4}} |
element initialization |
M = A(:) |
M = vectorise(A) |
|
M = [A B] |
M = join_horiz(A,B) |
|
M = [A; B] |
M = join_vert(A,B) |
|
M |
cout << M << endl or M.print("M =") |
|
A = randn(2,3) |
mat A = randn(2,3) |
|
B = randn(4,5) |
mat B = randn(4,5) |
|
F = {A; B} |
field<mat> F(2,1), F(0,0) = A or F(1,0) = B |
fields store arbitrary objects, such as matrices |