python - numpy array to permutation matrix -

np.array([1,2,3]) 

i've got numpy array. turn numpy array tuples of each 1:1 permutation. this:

np.array([     [(1,1),(1,2),(1,3)],     [(2,1),(2,2),(2,3)],     [(3,1),(3,2),(3,3)], ]) 

any thoughts on how efficiently? need operation few million times.

if you're working numpy, don't work tuples. use power , add dimension of size two. recommendation is:

x = np.array([1,2,3]) np.vstack(([np.vstack((x, x, x))], [np.vstack((x, x, x)).t])).t 

or:

im = np.vstack((x, x, x)) np.vstack(([im], [im.t])).t 

and general array:

ix = np.vstack([x _ in range(x.shape[0])]) return np.vstack(([ix], [ix.t])).t 

this produce want:

array([[[1, 1],         [1, 2],         [1, 3]],         [[2, 1],         [2, 2],         [2, 3]],         [[3, 1],         [3, 2],         [3, 3]]]) 

but 3d matrix, can see when looking @ shape:

out[25]: (3l, 3l, 2l) 

this more efficient solution permutations array size get's bigger. timing solution against @kasra's yields 1ms mine vs. 46ms 1 permutations array of size 100. @ashwinichaudhary's solution more efficient though.