How to selectively save the names and values of some workspace variables to a CSV file?

6 visualizzazioni (ultimi 30 giorni)
Consider the following
a = 3; b=7; c=5; d=9;
save('test.mat')
data = load('test.mat');
toSave = { 'b' ; 'c' };
writecell( horzcat( toSave , <missing code> ) ,'myfile.csv')
what should be the <missing code> in order to have the content of myfile.csv to be
b , 7
c , 5
?

Risposta accettata

Rik
Rik il 15 Set 2020
This should do the trick. If the variables aren't scalars the results might not be what you want.
a = 3; b=7; c=5; d=9;
save('test.mat')
data = load('test.mat');
toSave = { 'b' ; 'c' };
s=struct;
for n=1:numel(toSave)
s.(toSave{n})=data.(toSave{n});
end
writecell( horzcat( toSave , struct2cell(s) ) ,'myfile.csv')
  3 Commenti
Rik
Rik il 15 Set 2020
You could also ask load to only retrieve the variables of interest:
s=load('test.mat',toSave{:});
Note that loops have a pretty good performance, as long as you aren't dynamically growing arrays and there isn't a direct function to do it.

Accedi per commentare.

Più risposte (1)

Diaa
Diaa il 15 Set 2020
Thanks to the inspiring answer of @Rik, I found the following solution without the need to use for loop
a = 3; b=7; c=5; d=9;
save('test.mat')
data = load('test.mat');
toSave = { 'b' ; 'c' };
writecell(horzcat(toSave, cellfun( @(C) data.(C),toSave,'UniformOutput',false)),'myfile.csv')
  2 Commenti
Rik
Rik il 15 Set 2020
cellfun will only hide the loop. In general it will be slower that a well-written for loop. Just something to be aware of.
Diaa
Diaa il 15 Set 2020
You are right, but I can't resist minimizing the code despite knowing it is not the best practice :)

Accedi per commentare.

Prodotti


Release

R2020a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by