Searching and Manipulating Cell Array

I have a cell array A:

[3,2]

[2,4]

[4,5]

[2,5]

[6]

and cell array B:

[2,3,4]

[2,5]

[4,5,6]

In the array B I can see that [2,3,4] is made of union of [3,2] and [2,4] from A. [2,5] in B is made from union of only [2,5] from A. [4,5,6] in B is made of [4,5] and [6] in A.

So, I want to find the index of elements in A union of which make each array element of B. How do I do this?

Thanks guys

[Azzi Abdelmalek]

what are the sizes of elments in B? can they be >3. The same question about A

[Sam Da]

Yes they can be of any length. Thanks for interest. I think Matt's code already works for a general case. But if you have a simpler solution please do post.

Answer 1

[Matt J]

subsetmap=false(length(A), length(B));

for bb=1:length(B)

    for aa=1:length(A)

      subsetmap(aa,bb) = all(ismember(A{aa},B{bb}));

     end

end

%this next line is unnecessary if logical indexing is sufficient

indices=arrayfun(@(col) find(subsetmap(:,col)),1:length(B),'uni',0);

[Sam Da]

Thanks for quick reply. But can I make it so that it can consider union of more than 2 elements in A. i.e. there can be an element in B which is made of 1,2,3,4 all the way up to all elements of A.

[Matt J]

Yes. My code already does that.

[Matt J]

By the way. My code doesn't check to see whether all B{i} can actually be formed from the union of some A{j}. It only works if this is already the case. I leave it to you to add the appropriate error checks.