selecting the closest possible date

Hi ,

I have the following matrix:

1226  727198
1226  727563
1226  727929
1226  728294
1226  728659
1226  729024
1226  729390
1238  727198
1238  727563
1238  727929
1238  728294
1238  728659
1238  729024
1238  729390
1238  729755
1238  730120
1238  730485
1238  730851
1238  731216
1255  727198
1255  727563
1255  727929
1255  728294
1255  728659
1255  729024

where the first column is an ID and the second is date. I have a date X=728628.Is there a simple line or function that would enable me to select from each unique ID the date which is directly higher than X .i.e i want for each unique ID the date which is higher but closest to 728628. I can do that by writing many lines but i have the feeling that there is a very simple and direct way to do so.

[Jan]

It would be helpful if you explain the type of output you want. Do you want the concerning line indices related to the full matrix? Or do you want the index related to the IDs? Of course we can guess, what you could want. But an almost machting solution can be worse (wasting more programming time) than no solution.

[joseph Frank]

I just want to find the rows that match my criteria

[Oleg Komarov]

Can you specify what you mean by "directly higher"?

Answer 1

[Teja Muppirala]

If your matrix is M, and you have the Statistics Toolbox, then:

X = 728628;
grpstats(M(:,2),M(:,1),@(z)min(z(z > X)))

This returns

ans =

        728659
        728659
        728659

Which is the first number higher than 728628 for each of the 3 groups.