How to create matrix with Standard deviation from every 5 columns of another matrix?

I have matrix 100x20, how to create a new matrix with SD of evey 5 columns from the first one, to have 100x4 new matrix.

Answer 1

[the cyclist]

Here is one way:

% Pretend data [Use your actual data instead]
M = randn(100,20);

% Reshape into the slices needed for standard deviation
M3 = reshape(M,100,5,4);

% Take standard deviation along dimension 2
S3 = std(M3,0,2); % FYI, the 0 here means "use sample, not population std dev". (See documentation.)

% Reshape back to 2-d
S = reshape(S3,100,4);

Answer 2

[Walter Roberson]

M = rand(100,20);
Mstd5 = reshape(std(reshape(M, size(M,1), 5, []), [], 2), size(M,1), []);
size(Mstd5)

ans = 1×2

100 4