How to count the number of rows per group

Hi, I simply want to count the number of rows per group (for unique combinations of a, b and c). Groups where either a, b, or c are NaN should have a number of rows of NaN in the desired_output table.

% Original table
a = [1, 1, 1, 1, 2, 3, 3, 3]';
b = [NaN, NaN, NaN, NaN, 10, 23, 23, 23]';
c = [NaN, NaN, NaN, NaN, 5, 6, 6, 6]';
T = table(a, b, c)

T = 8×3 table
    a     b      c 
    _    ___    ___

    1    NaN    NaN
    1    NaN    NaN
    1    NaN    NaN
    1    NaN    NaN
    2     10      5
    3     23      6
    3     23      6
    3     23      6

% desired_output
a = [1, 1, 1, 1, 2, 3, 3, 3]';
b = [NaN, NaN, NaN, NaN, 10, 23, 23, 23]';
c = [NaN, NaN, NaN, NaN, 5, 6, 6, 6]';
d = [NaN, NaN, NaN, NaN, 1, 3, 3, 3]';
desired_output = table(a, b, c, d)

desired_output = 8×4 table
    a     b      c      d 
    _    ___    ___    ___

    1    NaN    NaN    NaN
    1    NaN    NaN    NaN
    1    NaN    NaN    NaN
    1    NaN    NaN    NaN
    2     10      5      1
    3     23      6      3
    3     23      6      3
    3     23      6      3

Thank you,

[Rik]

I don't fully understand how you determine d. Can you explain why the last elements are 1,3,3,3? And what exactly is the input data; T or a,b,c?

And what have you tried? This smells like homework to me.

[Blue]

Hi. I simply want the number of rows per unique combinations of a, b and c. If a, b or c are NaN then the number of rows should be NaN. There is 1 combination of 2, 10 and 5 for a, b and c which yields 1 in column d of desired output. No, this is not homework. I have tried going the groupcounts way then merging the results back in but this seems clumsy. Im sure there a one liner to do this.

[Image Analyst]

I tried some ways using table2array and unique but none of them was a one-liner. If you have a few lines of code that does it, just go with that. Sometimes longer code is better because it's more readable and understandable rather than a cryptic one-liner that no one can understand.

[Blue]

% Here is what I would attempt to go from T to desired output but Im stuck at the last row and anyhow this feels clumsy.

% Calculate number of rows per group

g = groupcounts(T, {'a', 'b', 'c'}, IncludeMissingGroups = false);

g.Percent = [];

g = renamevars(g, 'GroupCount', 'n_rows');

% Find unique combinations

[idx, idy] = ismember(T(:, {'a', 'b', 'c'}), g(:, {'a', 'b', 'c'}), 'rows');

% Assign back to T. Stuck here

% T = g(idy(idx), 4)];

Answer 1

[David Hill]

A=[a,b,c];
u=unique(A,'rows');
d=zeros(size(A,1),1);
idx=any(isnan(A),2);
d(idx)=nan;
f=find(~idx);
for k=1:length(f)
    d(f(k))=sum(ismember(A,A(f(k),:),'rows'));
end