MATLAB: Reshape array of cells with different column sizes into matrix

reshape

Hi. I need to reshape a T1 = [1×5] cell array, where each cell is a [5×1] cell array of numbers into a S1 = [m-by-n] matrix of numbers. The problem is the initial cell array T1 may have different number of rows in each cell – meaning I don't have a square shaped array to use " cell2mat " function.
Example:
T1=[1x5] cell array
T1{1}   T1{2}   T1{3}   T1{4}   T1{5}
 1       2       3       4       5
 6       7       8       9      10
11      12      13      14      15
16      17      18      19      20
21      22      23      24
I need to add all of the columns together:
T2=
 1   2   3   4   5
 6   7   8   9  10
11  12  13  14  15
16  17  18  19  20
21  22  23  24
And then reshape the array into [m-by-n] array (always square shaped), in this example S1=[3×8]:
S1=
 1  2  3  4  5  6  7  8
 9 10 11 12 13 14 15 16
17 18 19 20 21 22 23 24
I thought that maybe I could reshape T1 into [x-by-1] array and then reshape that array in to S1=[m-by-n]. But what should i do in order to get results like this:
T3=
1
2
3
4
...
24
And not like
1
6
11
16
21
2
...
24
Any help would be appreciated.

Best Answer

Here is a way:
T = {[1:5:21]',[2:5:22]',[3:5:23]',[4:5:24]',[5:5:20]'}
m=3
n=8
[M, tf] = padcat(T{:})
M = M.' 
out = reshape(M(tf.'),n,m).' % swap m and n, and then transpose
PADCAT concatenates unequal sized vectors by padding shorter ones with NaNs. This function can be downloaded from the File Exchange: https://uk.mathworks.com/matlabcentral/fileexchange/22909-padcat-varargin-
Related Question