Advice on Indexing A table.

Good Afternoon,

I have another probably easy to answer question. 

If I have a table like the one below. Is it possible to sort it so that for any number such as 38 here, I keep only the row with the earliest date and remove the rest? For        this example I would only have -- 38  '07/07/2014'? 

Noting too that the earliest date is not always the first entry. 

Any advice or videos to watch would be helpful and very much appreciated. 

Thank you!

    35                 '24/07/2014'    
     36                 '08/07/2014'    
     38                 '06/08/2014'    
     38                 '07/07/2014'    
     38                 '11/07/2014'    
     38                 '11/08/2014'    
     38                 '14/07/2014'    
     38                 '15/07/2014'    
     38                 '16/07/2014'    
     38                 '28/07/2014'    
     46                 '05/08/2014'    
     46                 '14/08/2014'

Answer 1

[Azzi Abdelmalek]

a={35                 '24/07/2014'    
 36                 '08/07/2014'    
 38                 '06/08/2014'    
 38                 '07/07/2014'    
 38                 '11/07/2014'    
 38                 '11/08/2014'    
 38                 '14/07/2014'    
 38                 '15/07/2014'    
 38                 '16/07/2014'    
 38                 '28/07/2014'    
 46                 '05/08/2014'    
 46                 '14/08/2014'}
c1=cell2mat(a(:,1));
c2=datenum(a(:,2),'dd/mm/yyyy');
[ii,jj,kk]=unique(c1);
min_date=accumarray(kk,c2,[],@(x) {datestr(min(x))});
out=[num2cell(ii) min_date]

Depending on what you mean by earliest, use min or max

Answer 2

[Peter Perkins]

Let's assume you have these in your workspace:

>> x = [35 36 38 38 38 38 38 38 38 38 46 46]';
>> date = {'24/07/2014' '08/07/2014' '06/08/2014' '07/07/2014' '11/07/2014' '11/08/2014' '14/07/2014' '15/07/2014' '16/07/2014' '28/07/2014' '05/08/2014' '14/08/2014'}';

Using R2014b or later, make a table, and convert the strings to datetimes (notice that it figured out that you meant day/month/year, and not month/day/year):

>> data = table(x,date);

>> data.date = datetime(data.date)
data = 
    x        date    
    __    ___________
    35    24-Jul-2014
    36    08-Jul-2014
    38    06-Aug-2014
    38    07-Jul-2014
    38    11-Jul-2014
    38    11-Aug-2014
    38    14-Jul-2014
    38    15-Jul-2014
    38    16-Jul-2014
    38    28-Jul-2014
    46    05-Aug-2014
    46    14-Aug-2014

If the data are actually in a file, there are other ways to do that, but let's assume workspace variables. Now sort the table, and find the first row for each unique value of x. That's your desired result.

>> data = sortrows(data);

>> [~,ui] = unique(data.x,'stable');
>> udata = data(ui,:)
udata = 
    x        date    
    __    ___________
    35    24-Jul-2014
    36    08-Jul-2014
    38    07-Jul-2014
    46    05-Aug-2014