Indexing a cell array according to another cell array

So I have two cell arrays, each represent onset and offset times of events, so they have the same size and they correspond to each other, for example:

A = {[2,3,4],[3,6],[5,7,9,10]};
B = {[2.5,4.2,4.7],[3.2,7.4],[6.2,7.6,9.4,11.3]}

I now was able to delete certain entries in A according to a different indexing vector c,let's say new A is (anything equals to 3 or between 7 to 9 is deleted):

A = {[2,4],[6],[5,10]}

How do I delete the corresponding entries in B? I want to achieve:

B = {[2.5,4.7],[7.4],[6.2,11.3]}

In other words delete the second, first, and second and third entires in the three cells in B. I tried to do it after entries were deleted from A, but the sizes and entries no longer match. It seems I have to somehow index A and B at the beginning before I mess with A, but I am not sure how to do this...

Answer 1

[Andrei Bobrov]

n = repelem(1:numel(A),cellfun(@numel,A));
B1 = [B{:}];
A1 = [A{:}];
t = A1 ~= 3 & (A1 < 7 | A1 > 9);
out = accumarray(n(t)',B1(t),[],@(x){x});

or

n = repelem(1:numel(A),cellfun(@numel,A))';
C = [n(:),[A{:}]',[B{:}]'];
C = C(C(:,2) ~= 3 & (C(:,2) < 7 | C(:,2) > 9),:);
[ii,jj] = ndgrid(C(:,1),1:2);
out = accumarray([ii(:),jj(:)],reshape(C(:,2:3),[],1),[],@(x){x});

Answer 2

[Azzi Abdelmalek]

A = {[2,3,4],[3,6],[5,7,9,10]};
B = {[2.5,4.2,4.7],[3.2,7.4],[6.2,7.6,9.4,11.3]}
idx=cellfun(@(x) x~=3 & ~(x<=9 & x>=7),A,'un',0)
C1=cellfun(@(x,y) x(y),A,idx,'un',0)
C2=cellfun(@(x,y) x(y),B,idx,'un',0)
celldisp(C1)
celldisp(C2)

[alicia che]

Thank you! This works! I have one other question and hope you could help...I am trying to use a logical vector as my constrain (instead of x=3 or x<=9 and x>=7), for example C = [0,0,1,0,0,0,1,1,1,0,0,0], how do I rewrite idx? Thank you so much!!!

[Azzi Abdelmalek]

idx for what?

[alicia che]

Sorry in your previous answer

idx=cellfun(@(x) x~=3 & ~(x<=9 & x>=7),A,'un',0)

I just wanted to change this so I can index according to C

[Azzi Abdelmalek]

And how to apply C to your cell array?

[alicia che]

A = {[2,3,4],[3,6],[5,7,9,10]}.

The elements in the cell array are time points in a time series from 0 to 10.

C = [0,0,1,0,0,0,1,1,1,0,0,0],

which contains 10 elements (same as the size of the time series). I want to remove any element in A (and corresponding ones in B) that corresponding to a "1" in C (x=3, or x>=7 and <=9). The result would be the same as your original solution:

A = {[2,4],[6],[5,10]}, 
B = {[2.5,4.7],[7.4],[6.2,11.3]}.

Thanks!

[Azzi Abdelmalek]

C does not contain 10 element but 12.

[alicia che]

I am sorry please ignore the last 2 zeros. C = [0,0,1,0,0,0,1,1,1,0]. Actually it doesn't matter too much in this case.

[Azzi Abdelmalek]

Ok, but A contains just 9 elements!

[alicia che]

I was using

A1=cellfun(@(x) x(~ismember(x,find(C))),A,'uni',false)

but it returns the final result I want but not index like your code did, so I can't apply it to B,and not sure how to fix it.

[Azzi Abdelmalek]

To make your question clear, post an example and explain clearly what you want, and post the expected result