## Rearrange a given array

### luca (view profile)

on 16 Sep 2019
Latest activity Commented on by luca

on 16 Sep 2019

### Guillaume (view profile)

Given a vector
V = [ 19 15 11 2 16 3 1 18 14 3 18 19 20 1 13 4 14 3 1 2 16 4 3 1 19 3 20 4 13 1 15 2 18 4 1 20 19 17 3 1 13 20 3 4 17 18 19 20 14 15 11 2]
I would like to rearrange V in this way.
every 10 element I want to collect the elements that are in the same column of A
A= [0 15 11 2 4 0 3 1 13;
14 0 16 0 0 0 0 0 0;
0 0 0 0 8 0 0 0 0;
0 0 18 0 0 0 0 0 0;
0 0 19 0 0 0 0 0 0;
0 0 0 0 0 0 0 0 0];
In this case, we want to collect 11,16,18,19 and also 4,8. Doesn't mind for the others that are just one in the columns, but if we have the same number I want to collect this two number together (see below with 3)
So for example, for the first 10 elements:
19 15 11 2 16 3 1 18 14 3
they will be rearrange as
19 11 16 18 15 2 3 3 1 14

R2019a

### Guillaume (view profile)

on 16 Sep 2019
Edited by Guillaume

### Guillaume (view profile)

on 16 Sep 2019

This would work as long as V has a multiple of 10 elements:
%demo data
V = [ 19 15 11 2 16 3 1 18 14 3 18 19 20 1 13 4 14 3 1 2 16 4 3 1 19 3 20 4 13 1 15 2 18 4 1 20 19 17 3 1 13 20 3 4 17 18 19 20 14 15];
A= [0 15 11 2 4 0 3 1 13;
14 0 16 0 0 0 0 0 0;
0 0 0 0 8 0 0 0 0;
0 0 18 0 0 0 0 0 0;
0 0 19 0 0 0 0 0 0;
0 0 0 0 0 0 0 0 0];
assert(mod(numel(V), 10) == 0, 'V length is not a multiple of 10');
V10 = reshape(V, 10, []); %rearrange V in batches of 10
[~, where] = ismember(V10, A); %find location of elements in A
[~, col] = ind2sub(size(A), where); %convert location into column index
[~, order] = sort(col, 1); %and sort by column order
newV = reshape(V10(order + (0:size(order, 1):numel(order)-1)), 1, []) %and rearrange columns of V10 according to that order
Note that elements of V not found in A get reordered first in each group of 10.

luca

### luca (view profile)

on 16 Sep 2019
The fact to insert NaN complicate and create confusion once we have to reorder. If you have other suggestion on how to do it, please let me know!
Thanks
Guillaume

### Guillaume (view profile)

on 16 Sep 2019
Oh, yes, forgot I'm not sorting the numbers but their column position (which is 0 for NaNs).
newV = V(~isnan(V))
for the last line. The rest as before.
luca

on 16 Sep 2019