Assign a Sub Array to Array Knowing the Number of Dimensions at Run Time

4 ビュー (過去 30 日間)
Royi Avital
Royi Avital 2022 年 2 月 10 日
コメント済み: Royi Avital 2022 年 2 月 11 日
Assume we have tA and tB with the same number of dimesions. We also have all(size(tB) <= size(tA)) == true.
The task is to embed tB in tA. Something like: tA(1:size(tB, 1), 1:size(tB, 2), ..., 1:size(tB, n)) = tB. Yet since we know the number of dimensions only at runtime, it can't be written explicitly like that.
The question, is there an efficient way to do so without eval or explicitly computer the cartesian product and use sub2ind()?

採用された回答

Royi Avital
Royi Avital 2022 年 2 月 10 日
編集済み: Royi Avital 2022 年 2 月 11 日
OK, It turns out it can be done using Cell Arrays:
vSizeB = size(tB);
numDims = length(vSizeB); %<! Equals to ndims(tB)
cIdx = cell(numDims, 1);
for ii = 1:numDims
cIdx{ii} = 1:vSizeB(ii);
end
tA(cIDx{:}) = tB;
  6 件のコメント
Stephen23
Stephen23 2022 年 2 月 11 日
編集済み: Stephen23 2022 年 2 月 11 日
"Your code won't work as you need the 1:vSizeB(ii) vector and not only a single number"
Lets try it:
tA = nan(5,4,3);
tB = randi(9,4,3,2);
tmp = arrayfun(@(n)1:n,size(tB),'uni',0);
tA(tmp{:}) = tB
tA =
tA(:,:,1) = 4 9 5 NaN 2 1 5 NaN 2 2 2 NaN 7 5 4 NaN NaN NaN NaN NaN tA(:,:,2) = 6 1 8 NaN 9 4 5 NaN 5 4 4 NaN 4 3 3 NaN NaN NaN NaN NaN tA(:,:,3) = NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
Royi Avital
Royi Avital 2022 年 2 月 11 日
@Stephen, Well you edited the comment. It was, at first, only num2cell().

サインインしてコメントする。

その他の回答 (1 件)

Rik
Rik 2022 年 2 月 10 日
There is probably a better way, but you can fill a cell array with the index vectors (simple loop with ndims), and then use this:
tA(inds{:}) = tB;

カテゴリ

Help Center および File ExchangeMatrices and Arrays についてさらに検索

製品

Community Treasure Hunt

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

Start Hunting!

Translated by