Save multiple functions on a single vector

I have a function with 3 outputs

[out1,out2,out3=myfunction(dataset); (every output is a single number , it is the min avg and maximum value of the dataset)

I need to use a loop, so I will have out1,out2,out3 for every day and I want to save them in a matrix,I want every row to be a vector with the 3 outputs

Any suggestions?

[Guillaume]

How is every day defined? Is the day a variable of the dataset?

[kounoupaki87]

The dataset is a 3D matrix (lon, lat, date)

[kounoupaki87]

You can find a sample attached if helps :D

[Guillaume]

What version are you on that you're still using nanmean, nanmax, nanmin?

Despite your statement, yout minmaxavg function doesn't return scalar outputs, but vectors. Note that the function can be greatly simplified but we'll worry about that once I've understood what it is you want.

I'm still unclear what it is you want. How is a day defined? Is each page of the matrix, a day, an hour, something else? Maybe another array defined what a page corresponds to. Also, do the rows and columns have any meaning, your current code takes the min/max/mean of all the elements, so it doesn't appear that the row/column separation is important.

[kounoupaki87]

I have a different dataset for every day (Its temperature values for Longitude and latitude points and I have them combined in a 3D matrix (lon, lat,day), now the row-column separation is not important, I just want the mim avg and max temperature values of the domain

Answer 1

[Guillaume]

"I have a different dataset for every day (Its temperature values for Longitude and latitude points and I have them combined in a 3D matrix (lon, lat,day)"

"I will have out1,out2,out3 for every day and I want to save them in a matrix,I want every row to be a vector with the 3 outputs"

I am a bit confused then because that's what your minmaxavg function already do. However, as I said, it can be greatly simplified.

If you have R2018b or later:

dailymean = squeeze(mean(your3Dmatrix, [1 2], 'omitnan'));
dailymax = squeeze(max(your3Dmatrix, [], [1 2], 'omitnan'));
dailymin = squeeze(min(your3Dmatrix, [], [1 2], 'omitnan'));

Otherwise, if you have R2015a or later:

dailymean = mean(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), 1, 'omitnan');
dailymax = max(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), [], 1, 'omitnan');
dailymin = min(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), [], 1, 'omitnan');

And if you're still stuck on a version prior to R2015a, then you need the stats toolbox:

dailymean = nanmean(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), 1);
dailymax = nanmax(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), [], 1);
dailymin = nanmin(reshape(your3Dmatrix, [], size(your3Dmatrix, 3)), [], 1);

In any case, you don't need a loop.

Answer 2

[Jesus Sanchez]

One possibility is to work with them as vectors and then unite them in one matrix.

everyDay = 10; % N
out1 = zeros(everyDay,1); % Size Nx1
out2 = zeros(everyDay,1);
out3 = zeros(everyDay,1);
for n=1:everyDay
    [out1(n),out2(n),out3(n)]=myfunction(dataset);
end
savedData = [out1(n),out2(n),out3(n)];