MATLAB: Adding an id for similar rows

indexrowsunique

Hello,
I have a data set with row and column numbers in reference to a grid, and then a value for that cell.
Example of data set (Attached .csv titled "input")
row  col  value;
124  95  7983;
124  96  16142;
124  97  94846;
124  98  4557;
124  95  8935;
124  96  496746;
124  97  352734;
124  98  25374;
124  95  790574;
124  96  346;
124  97  3745;
124  98  19774;
What I'd like to do is add a column and for all the first unique row/column combinations, and populate the new column with 1. And then for the next repetition of those row/column combinations, use a 2 and so on. So this example data set would look like the following
ID  row  col  value;
1  124  95  7983;
1  124  96  16142;
1  124  97  94846;
1  124  98  4557;
2  124  95  8935;
2  124  96  496746;
2  124  97  352734;
2  124  98  25374;
3  124  95  790574;
3  124  96  346;
3  124  97  3745;
3  124  98  19774;
I figured I could use the unique function to accomplish this, but I've been unsuccessful thus far. Does anyone have an easy way in mind to do this? Thanks so much. Oh, and the actual data files I'll be trying to apply this two are around 7 million lines, with around 600 different repetitions of the same row/column combinations, but different values going with the combinations each time. Thanks so much for your time!

Best Answer

One way to do it:
data = [124 95 7983; 124 96 16142; 124 97 94846; 124 98 4557; 124 95 8935; 124 96 496746; 124 97 352734; 124 98 25374; 124 95 790574; 124 96 346; 124 97 3745; 124 98 19774];
%prepare array to receive ids:
ids = zeros(size(data, 1), 1);   
%find out where the unique rows of data are distributed:
[~, ~, uid] = unique(data(:, [1 2]), 'rows'); 
%group together the row indices with the same uid:
grouprows = accumarray(uid, 1:size(data, 1), [], @(v) {sort(v)});
%for each cell in grouprow create array 1,2,3... of final ids:
groupids = cellfun(@(ix) (1:numel(ix))', grouprows, 'UniformOutput', false);
%fill up ids using grouprows as indices and groupids as values:
ids(vertcat(grouprows{:})) =  vertcat(groupids{:});
%append ids to existing data:
newdata = [ids data]
Related Question