MATLAB: How can I reshape vector data into a matrix

matrixreshaping data

Can anyone help me with this please… I have a data set, shaped so. This needs to be graphed in 3D.

The first column represents locations on X-axis, the 2nd column – locations on Y-axis, and the third column is my data. The 1st & 2nd column make a grid.

How do I reshape this to make the 3rd column into a 2×2 matrix, so that the whole thing can be graphed as a surface?

Thank you beforehand!

Best Answer

How do I reshape this to make the 3rd column into a 2x2 matrix, so that the whole thing can be graphed as a surface?

You don't, it's completely unnecessary

surf(yourmatrix(:, 1), yourmatrix(:, 2), yourmatrix(:, 3));

If you wanted to do the reshape (which again is completely unnecessary), and assuming that the X and Y coordinates that are supposed to be identical are actually identical down to the last binary digit, then:

[~, ~, col] = unique(yourmatrix(:, 1));
[~, ~, row] = unique(yourmatrix(:, 2));
X(row, col) = yourmatrix(:, 1);
Y(row, col) = yourmatrix(:, 2);
Z(row, col) = yourmatrix(:, 3);

Related Solutions

MATLAB: Can this silly accumarray for-loop be removed by a vectorization

A = randi(32280,28);
[UA,~,idx] = unique(A(:,1));
[X,Y] = ndgrid(idx, 2:size(A,2));
newcolumns = accumarray([X(:), Y(:)], reshape(A(:,2:end),[],1), [], @mean);

The ndgrid needs to start in the second column and only extract the 2nd through end columns of A for vals.

MATLAB: For loop using unique values

Indeed using find would be madness. For the record, find is over used by beginners and is rarely needed. For that matters, for loop are also overused and are also not needed in this case.

There are many ways to do this:

Using groupsummary:

maxdischarge = groupsummary(yourmatrix(:, 4), yourmatrix(:, 2), 'max');

Using splitapply:

[group, month] = findgroups(yourmatrix(:, 2));  %may not be needed if the month column is integers from 1 to 12
maxdischarge = splitapply(@max, yourmatrix(:, 4), group);  %maxdischarge(i) corresponds to month(i)

Using old fashioned accumarray:

[month, ~, group] = unique(yourmatrix(:, 2));  %works like findgroups, with different order of output. Again may not be needed
maxdischarge = accumarray(yourmatrix(:, 4), group, [], @max);

Or even using a loop:

maxdischarge = zeros(12, 1)
for month = 1:12
    maxdischarge = max(yourmatrix(yourmatrix(:, 2) == month, 4));
end

Note that I would recommend that you convert your matrix into a table:

t = array2table(yourmatrix, 'VariableNames', {'Day', 'Month', 'Year', 'Discharge'});

In which case, you definitively should use groupsummary:

groupsummary(t, 'Month', 'max', 'Discharge');  %calculate max of discharge grouped by month

Even better would be to convert the first 3 columns into a datetime array:

t = table(datetime(yourmatrix(:, 3), yourmatrix(:, 2), yourmatrix(:, 1)), yourmatrix(:, 4), 'VariableNames', {'Date', 'Discharge'});
groupsummary(t, 'Date', 'monthofyear', 'max', 'Discharge')

Best Answer

Related Solutions

MATLAB: Can this silly accumarray for-loop be removed by a vectorization

MATLAB: For loop using unique values

Related Question