I have a=[[1 2 ... 3][4 5 ... 6]...[7 8 ... 9]].
I need a=[[[1 1 1] [2 2 2] ... [3 3 3]][[4 4 4] [5 5 5] ... [6 6 6]]...[[7 7 7] [8 8 8] ... [9 9 9]]]
I basically need each element in a to become a tuple of 3 values of itself.
Add a comment
|
1 Answer
Tile 3 times on a columnar version and finally map to tuples, like so -
map(tuple,np.tile(a.ravel()[:,None],(1,3)))
If you are looking for a 3D array as listed in the expected output in the question, you could do -
np.tile(a[:,:,None],(1,1,3))
Sample run -
In [32]: a
Out[32]:
array([[1, 2, 3],
[4, 5, 6],
[7, 8, 9]])
In [33]: map(tuple,np.tile(a.ravel()[:,None],(1,3)))
Out[33]:
[(1, 1, 1),
(2, 2, 2),
(3, 3, 3),
(4, 4, 4),
(5, 5, 5),
(6, 6, 6),
(7, 7, 7),
(8, 8, 8),
(9, 9, 9)]
In [34]: np.tile(a[:,:,None],(1,1,3))
Out[34]:
array([[[1, 1, 1],
[2, 2, 2],
[3, 3, 3]],
[[4, 4, 4],
[5, 5, 5],
[6, 6, 6]],
[[7, 7, 7],
[8, 8, 8],
[9, 9, 9]]])
9 Comments
Sibi
Is there any way to get square brackets and no commas? :p
Divakar
@Sibi Use without map :
np.tile(a.ravel()[:,None],(1,3)).Divakar
Are you dealing with
np.matrix or np.array?Divakar
Well then most probably you are just confusing yourself with the printing of arrays, which has nothing to do with how outputs are generated. The commas are just to show different elements of the input. Did you even try the code?
Divakar
@Sibi because that still is not a NumPy array. Do
a = np.asarray(a) and try again. |