# Sort matrix based on unique values in one column

838 views (last 30 days)
Ida Rosendahl on 14 Apr 2016
Commented: IVAN JIMENEZ on 9 Mar 2021
Hi guys. I can't seem to overcome my problem. I have a matrix with 5 columns and a lot of rows. I need to sort the whole matrix based on unique values in column 1. How do I do that? So I need to find the unique values (its a timestamp from my experiment) in column one and also the corresponding values in the rest of the columns and everything else needs to be deleted. Thanks!

Stephen on 14 Apr 2016
Edited: Stephen on 14 Apr 2016
[~,idx] = sort(mat(:,1)); % sort just the first column
sortedmat = mat(idx,:); % sort the whole matrix using the sort indices
You can even do the same thing with unique to get the rows corresponding to unique values in the first column:
[~,idu] = unique(mat(:,1))
uniquerows = mat(idu,:);
Note that unique has options for controlling the order of its outputs, and whether it takes the first or last matching element. You should check its documentation and pick the best options for you requirements.
IVAN JIMENEZ on 9 Mar 2021
you saved my life, thnks

Guillaume on 14 Apr 2016
I'm not sure what you mean by "and everything else needs to be deleted". Does that means that rows with identical first column also have the exact same values for all the other columns, or that you want to somehow merge these identical rows into just one?
Assuming all duplicate rows are identical (but then why would you say unique values in one column?):
b = unique(a, 'rows');
Assuming you want to just keep one arbitrary row out of the ones that have identical first column:
[~, rows] = unique(a(:, 1));
b = a(rows, :);
If you want to somehow merge the duplicate rows, it probably can be done but you need to give more details.
STAVRINI TSANGARI on 23 Oct 2020
how can I get the average of the two duplicate columns?

Mohammad Bhat on 15 Feb 2018
sir, If we want to keep only maximum value besides second unique column value...