MATLAB: How to do pivot from table-type variable

pivotingtablevectorization

Hi,
I have this table-type data
tdata.yearID=[1950;1950;1950;1951;1951;1951;1951;1952;1952;1952];
tdata.cluster=[1;3;4;1;2;3;4;1;2;3];
tdata.GroupCount=[5;2;6;2;2;4;2;1;2;2];
tdata= struct2table( tdata)
tdata =
10×3 table
    yearID    cluster    GroupCount
    ______    _______    __________
     1950        1           5     
     1950        3           2     
     1950        4           6     
     1951        1           2     
     1951        2           2     
     1951        3           4     
     1951        4           2     
     1952        1           1     
     1952        2           2     
     1952        3           2
Currently I am using this code below to do pivot, but it was very slow for large data. Is there a way to vectorize this code?
[uniqYr, ~, JYr]= unique( tdata.yearID);
nEventData.yearID= [1: 2000]';
[ nEventData.cnt]= zeros( length(nEventData.yearID), 4);
for runYr= 1: length( uniqYr)
    indxYr= uniqYr( runYr);
    tloc  = (JYr== runYr);
    tt    = tdata( tloc, :);
    nEventData.cnt( indxYr, tt.cluster')= tt.GroupCount';        
end % for runYr
So results look like this for rows 1949-1955. Other rows have zeros.
nEventData.cnt (1949:1955,:)
ans =
     0     0     0     0
     5     0     2     6
     2     2     4     2
     1     2     2     0
     0     0     0     0
     0     0     0     0
     0     0     0     0

Best Answer

The output you described looks somewhat like the first example in the documentation for the unstack function. I think what you want is something close to the pivotedTdata variable created by this example:
tdata.yearID=[1950;1950;1950;1951;1951;1951;1951;1952;1952;1952];
tdata.cluster=[1;3;4;1;2;3;4;1;2;3];
tdata.GroupCount=[5;2;6;2;2;4;2;1;2;2];
tdata= struct2table( tdata)
pivotedTdata = unstack(tdata, 'GroupCount', 'cluster')
You can retrieve data for a particular year this way:
pivotedTdata(pivotedTdata.yearID == 1950, :)
Related Question