How to append new values from a loop to the bottom of an existing matrix?

Hi everyone - I have a 1x5 array of numbers ranging from 0-1 which is populated from elsewhere in the code from a loop (1st trial might produce 0, 0.4, 0.55, 0.6, 0.4, 2nd trial might produce 0.2, 0.4, 0.4, 0.5, 0.8)

But I want to create a new matrix which stores all the values from this loop as one matrix such that the 1st trial (0, 0.4, 0.55, 0.6, 0.4) is stored on the first row and the 2nd trial (0.2, 0.4, 0.4, 0.5, 0.8) is stored on the 2nd row etc so each new trial populates the new trial values onto a new row of the matrix

The closest I've got so far is:

%Number of trials
trials = 2;
for N = 1:trials
    all = [];
    all(N,:) = trial_values; 
end

Obvs this overwrites the pre-existing values when it goes to the next trial but places them on the correct rows, would anybody be able to help point out the bit I'm missing please so that it keeps the values from the previous trial?

[Alexander Halbleib]

Edit: Whoops - I just commented. Moving this to the "Answer" tab

[Jack Williams]

Perfect that's amazing - all working now! Appreciate the help :)

Answer 1

[Dyuman Joshi]

Since the data here is quite small, looping twice with 5 elements each, you can make do with pre-allocating as an empty matrix, as done below -

%Number of trials
trials = 2;
all=[];
for N = 1:trials
    all = [all trial_values];
end

In the above method, the array is growing with every iteration and if the number of iterations and/or the growing length are high, the whole process can easily become slow and is inefficient.

The recommended method is to pre-allocate a zeros matrix, in case of numeric data, according to the final size of the output (Read more - Pre - allocation)

trials = 2;
%pre-allocation
%use n = numel(trial_values) for a general case
%which is 5 in this case
all = zeros(1,trials*n);
for k = 1:trials
    %data entries will be 1-5, 6-10, 11-15, 16-20
    %so we use the loop index accordingly
    all(1,(5*(k-1)+1):5*k) = trial_values;
end

Answer 2

[Alexander Halbleib]

Hi Jack,

Your issue lies in the following line:

all = [];

That line resets the whole array to an empty array, and then your assignment adds creates an empty array of size N x length(trial_values), and inserts your latest results.

Instead, you should do something like the following:

trials = 2;
all = [];
for N = 1:trials
    all(N,:) = trial_values;
end

The downside here is that it's really inefficient. Every loop, Matlab has to create a new array for "all", because it's growing. There are a couple of ways to make it much faster!

The easiest way is to initialize your "all" array to be the right size:

trials = 2;
all = NaN(trials,length(trial_values));
for N = 1:trials
    all(N,:) = trial_values;
end

That way, matlab doesn't have to recreate an array and copy the relevant data over every loop!

Best of luck!