# is there a way to perform this task w/o using loops?

1 view (last 30 days)
Jeffrey on 29 Nov 2023
Commented: Jeffrey on 2 Dec 2023
% this script sets up a structure p. In this minimal version, p only contains 1 variable.
% then expands it into a 1x4 structure array q
% and inserts values into one of the variables. (Only the L variable is shown here.)
% and then retrieves the values
% The variable is a row vector and has name given by varName
% The values go into (and are retrieved from) the 2nd element of L because varIndex = 2;
p.varName = 'L';
p.varIndex = 2;
p.L = [3, 4, 5];
nval = 4;
q = repmat(p,1,nval);
vals = [6 7 8 9];
% Insert values
for i=1:nval
q(i).(p.varName)(p.varIndex) = vals(i);
end
Xv = zeros(1,nval);
% Retrieve values
for i=1:nval
Xv(i) = q(i).(q(1).varName)(q(1).varIndex);
end
##### 3 CommentsShow 1 older commentHide 1 older comment
Walter Roberson on 29 Nov 2023
Xv = arrayfun(@(IDX) q(IDX).(q(1).varName)(q(1).varIndex), 1:nval);
The assignment is more difficult to arrange.
Jeffrey on 2 Dec 2023
Understood. Thank you.

Walter Roberson on 29 Nov 2023
Edited: Walter Roberson on 29 Nov 2023
The more general vectorized strategy would probably be to set up cell arrays of values, and then use the fact that when you struct() and pass a cell array, then the output is a struct array the size of the cell array.
Jeffrey on 2 Dec 2023
Understood. I like this answer the best because it's elegant, but will probably stick with the loop instead of introducing cell arrays into my code, to keep it simple.

Matt J on 29 Nov 2023
Edited: Matt J on 29 Nov 2023
There are ways to do it without loops, but doing it without loops will be of no benefit to you. It will take more lines of code, consume more memory, and run more slowly.
p.varName = 'L';
p.varIndex = 2;
p.L = [3, 4, 5];
nval = 4;
q = repmat(p,1,nval);
vals = [6 7 8 9];
% Insert values
TMP=vertcat(q.(p.varName));
TMP(:,p.varIndex)=vals;
TMP=arrayfun(@(i)TMP(i,:), 1:nval,'uni',0);
[q.(p.varName)]=deal(TMP{:});
q.L
ans = 1×3
3 6 5
ans = 1×3
3 7 5
ans = 1×3
3 8 5
ans = 1×3
3 9 5
Jeffrey on 2 Dec 2023
Understood. Thank you!

Bruno Luong on 29 Nov 2023
Edited: Bruno Luong on 29 Nov 2023
You might consider store in table instead
NOTE: using table would be more convenient to access data, not necessary faster
p.varName = "L";
p.varIndex = 2;
nval = 4;
q = repmat(p,1,nval);
vals = [6 7 8 9];
% Insert values
for i=1:nval
q(i).(p.varName)(p.varIndex) = vals(i);
end
T = struct2table(q)
T = 4×3 table
varName varIndex L _______ ________ ______ "L" 2 0 6 "L" 2 0 7 "L" 2 0 8 "L" 2 0 9
T.varName(1)
ans = "L"
T.(T.varName(1))
ans = 4×2
0 6 0 7 0 8 0 9
T.(T.varName(1))(:,T.varIndex(1))
ans = 4×1
6 7 8 9
Jeffrey on 2 Dec 2023
Very good. Thank you. I hadn't heard of tables.

### Categories

Find more on Resizing and Reshaping Matrices in Help Center and File Exchange

R2023a

### Community Treasure Hunt

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

Start Hunting!