Create arrays of observations

1 view (last 30 days)
CARLOTTA DENTICO on 4 Aug 2021
Commented: Rik on 4 Aug 2021
Hello everybody,
Probably it is a very silly question but I am struggling on it..
I have a series of observations saved as 'pt_mean_lev1_1' up to 'pt_mean_lev1_348' (see picture below) and I want to make an array of them being a = 1x348. How can I do that?
Thank you very much!
3 CommentsShow 1 older commentHide 1 older comment
CARLOTTA DENTICO on 4 Aug 2021
These are monthly mean of temperature spatially averaged over a specific area of the world, for every months for 29 years (I have a total of 348 means i.e 12x29) that I have extracted from a NetCDf file..
I need these means since I need to calculate time series of winter months and summers months and calculate the mean of winter months in the 29 years and mean of summer months.
All that is to calculate the anomalies
CARLOTTA DENTICO on 4 Aug 2021
fram = (lon>=0 & lon<=20 & lat>=76 & lat<=80);
for k = 1:348
eval(['pt_lev1_' ,num2str(k), ' = reshape(pt_lev1(:,:,',num2str(k),'),[],1);']);
eval(['pt_mean_lev1_', num2str(k), '= mean(pt_lev1_',num2str(k),'(fram,:),1,''omitnan'');']);
end

Sulaymon Eshkabilov on 4 Aug 2021
Here is one of the viable solutions with dynamically naming variables and attaining their values from the workspace.
a = [];
for ii=1:348
Var = strcat('pt_mean_lev1_', num2str(ii));
a=[a, eval(Var)];
end
3 CommentsShow 1 older commentHide 1 older comment
Sulaymon Eshkabilov on 4 Aug 2021
Rik on 4 Aug 2021
In the screenshot you can see the number of elements of each element. Growing the array dynamically like this (instead of pre-allocating) results in terrible performance. I would also refrain from teaching people to use eval as the solution. As already evidenced in this thread, that is a source of trouble.
k=500000;N=100;%N=348 takes mare than the 55 second limit
tic
a = [];
for ii=1:N
a=[a, rand(k,1)];
end
toc
Elapsed time is 15.427005 seconds.
tic
a = zeros(k,N);
for ii=1:N
a(:,ii)=rand(k,1);
end
toc
Elapsed time is 0.513869 seconds.

CARLOTTA DENTICO on 4 Aug 2021
Do you mean here?
Basically, these are monthly mean of temperature spatially averaged over a specific area of the world, in every months for 29 years (I have a total of 348 means i.e 12x29) that I have extracted from a NetCDf file..
I need these means since I need to calculate time series of winter months (from October to April) and summers months (from May to September) and calculate the mean of winters and summers over 29 years.
All that is to calculate the anomalies of a single month compared to winters in the whole period..
Sulaymon Eshkabilov on 4 Aug 2021
I meant a separate thread and not here. Two questions in one thread that is confusing.

Rik on 4 Aug 2021
You should not name your variables dynamically. Why are you not storing them in an array when reading? What is wrong with the code below?
fram = (lon>=0 & lon<=20 & lat>=76 & lat<=80);
pt_lev1_=zeros(size(pt_lev1,1)*size(pt_lev1,2),348);
for k = 1:348
pt_lev1_(:,k) = reshape(pt_lev1(:,:,k),[],1);
end
pt_mean_lev1_=mean(pt_lev1_(fram,1:k),1,'omitnan')