matlab rows to column labels

Question

If I have data organized like this:

a c 1
a d 2
b c 3
b d 4

I want to use a and b as my row identifiers and c and d as my column identifiers, yielding a table like this:

  c d
a 1 2
b 3 4

I can accomplish this with a loop in the following manner:

old1 = {'a';'a';'b';'b'};
old2 = {'c';'d';'c';'d'};
dataa = [1 2 3 4]';

tabul = array2table(nan(2,2));
tabul.Properties.RowNames={'a','b'};
tabul.Properties.VariableNames={'c','d'};


for t = 1 : 4

    find1 = strcmp(old1(t),tabul.Properties.RowNames);
    find2 = strcmp(old2(t),tabul.Properties.VariableNames);
    tabul{find1,find2} = dataa(t);

end

Is there a more elegant way to solve this problem? Possible candidates: groupstats? splitapply?

score 3 · Accepted Answer · answered Sep 21 '15 at 22:46

3

You can use array2table to directly set the table values, assuming they are sorted (row-major) as in your example.

dataa = [1 2 3 4].';
tabul = array2table(reshape(dataa,2,2).');
tabul.Properties.RowNames={'a','b'};
tabul.Properties.VariableNames={'c','d'};

Also, you can pass the row and variable names directly to array2table:

dataa = [1 2 3 4].';
tabul = array2table(reshape(dataa,2,2).', 'RowNames',{'a','b'}, 'VariableNames', {'c','d'});

Either of the above gives the same result as your example, that is,

tabul = 
         c    d
         _    _
    a    1    2
    b    3    4

answered Sep 21 '15 at 22:46

Luis Mendo

110,752
13
76
147

This is good only as far as there are all the possible combinations between {'a','b'} and {'c','d'} in the starting dataset, however. When this is not the case, the do a transposition would require to create missing data points to get a rectangular data matrix to reshape. As far as I know, there is still no 'pivot table' command built in in matlab, but googling it is possible to find around some user package (never tried them however). – Giuseppe Nov 27 '15 at 19:26

matlab rows to column labels

1 Answers1