How to extract last element from the lists

Suppose if my List has

list = {[1,2],2,[],[1,2,3]}
eg:

list{1} = [1,2] = %% print last element ==>> 2 list{2} = 2 list{3} = [] list{4} = [1,2,3] ==> 3 new_list = {2,2,[],3}

I need to extract last element from each list{i} Output should be like

new_list = {2,2,0,3}

[dpb]

For what definition of nth element, pray tell? You have vectors of numel() = 2,1,0,3 for the cells and the ordinal position of your proposed solution is 1,1,[],3 which is, if anything, consistent in there is no consistency observable as to which is to be recalled (other than the third with no alternative available).

[Aswin Sandirakumaran]

list{1} = [1,2] = %% print last element ==>> 2 list{2} = 2 list{3} = [] list{4} = [1,2,3] ==> 3 new_list = {2,2,[],3}

[dpb]

That's not what the original said; no fair changing the problem and then acting like didn't... :)

[Aswin Sandirakumaran]

Oh common.. Yes I made a mistake to interpret ! But cuz of you.. I realised my mistake and changed it . And one more thing even before changing it someone understood my problem and he solved it.

Answer 1

[Walter Roberson]

Nth = @(L, N) L(N);
PadN = @(L,N) [L,zeros(1,N)];

list = {[1,2],2,[],[1,2,3]};
N = 2;
cellfun(@(L) Nth(PadN(L,N),N), list)

This is what you asked for, the nth element from the lists. But the example output you gave is asking for the last element of the list, not the n'th element.

Last = @(L) L(end);
NonEmpty = @(L) [L,zeros(1,1-length(L))];
cellfun(@(L) Last(NonEmpty(L)), list)

[Aswin Sandirakumaran]

Thanks ! Can you explain me how it is done?

Last = @(L) L(end);
NonEmpty = @(L) [L,zeros(1,1-length(L))];
cellfun(@(L) Last(NonEmpty(L)), list)

[Walter Roberson]

I need to run some errands for a while; perhaps someone else will have time to explain it.

[Walter Roberson]

You have a list of values in a cell, so it is easiest to run a cellfun() operation to calculate the results.

Now, if you knew ahead of time that the entries were all non-empty then you could just code

cellfun(@(L) L(end), list)

to take the last entry of each of them. But you do have empty entries, and you have defined that those entries should return 0. So we need to detect empty entries and pad them with a 0 before taking the last element of them.

One of the ways to pad with 0 conditionally is to concatenate with zeros(1,1-length(L)) . When the entry is empty then length() is 0 and 1-length() is 1-0 so that becomes zeros(1,1) which is a single 0 concatenated on to the end of the emptiness, leaving a 0. When the entry is not empty then length() is 1 or more and 1-length() is 0 or negative, so that becomes zeros(1,0 or less), which is emptiness concatenated to the end of the non-empty vector, leaving the vector unchanged.

You would be tempted to put those two functionalities together in a single cellfun, like

cellfun(@(L) [L,zeros(1,1-length(L))](end), list)   %WRONG

to try to do the concatenation and then take the last element of what results. However, MATLAB does not permit that syntax.

There is an actual function, subsref, that you can call to do indexing. But it is ugly to call and hard to understand. So it is easier to add in an auxillary function such as

Last = @(L) L(end)

and then you can

cellfun(@(L) Last([L,zeros(1,1-length(L))]), list)

The code I posted just broke this into one more level of operation to add a bit of clarity, doing the padding in another auxiliary function.