Grouping rows on the first column and obtain minimum value on second column

I have the following matrix A and would like to group it on the first column

A = [13 7
13 6
13 5
13 6
12 7
12 5];

and then obtain minimum value on second column to obtain matrix B

B = [13 5
12 5];

Could anyone give a hint?

Thanks!

Answer 1

[Torsten]

What do you mean by "group on the first column" ?

A = [13 7
    13 6
    13 5
    13 6
    12 7
    12 5];
A = sortrows(A,2,'ascend');
B = A(1:nnz(A(:,2)==A(1,2)),:)

B = 2×2

13 5 12 5

[H_W]

find minimum value in second column that has the same value in first column

[Adam Danz]

I interpreted the task as finding the minimum values in col 2 for each group in col 1. This solution returns rows of A that match the minimum value in col 2. @H_W, perhaps Torsten correctly interpreted your question but you might want to verify.

For example,

A = [13 7
  13 6
  13 5
  13 6
  12 4 %  changed from 7 to 4
  12 5];
A = sortrows(A,2,'ascend');
B = A(1:nnz(A(:,2)==A(1,2)),:)

B = 1×2

12 4

[H_W]

Thanks! I want to keep the first column, so in this example, output that I want is

B = 2×2
    13     5
    12     4

[Torsten]

So for all distinct values in the first column, you want to get one pair where the value in the second column is minimum ? Is it necessary to group the first column first (e.g. sort in ascending/descending order) or is this already done in A ?

[Adam Danz]

If I'm understanding correctly, that's what the solutions do in my answer.

Answer 2

[Adam Danz]

A = [13 7
13 6
13 5
13 6
12 7
12 5];

B = groupsummary(A,A(:,1),'min')

B = 2×2

12 5 13 5

Note that rows are sorted according to the grouping column values.

To keep the row order stable there are several options but I lean on arrayfun which is less readable than other options.

B = cell2mat(arrayfun(@(g){min(A(A(:,1)==g,:))}, unique(A(:,1),'stable')))

B = 2×2

13 5 12 5