Sort a table with different orders for different columns

Assume table A:

A = [
    'k' 2 4
    'a' 3 8
    'a' 4 5
    'k' 2 2
    'a' 3 16
    'k' 9 3
    'k' 9 8
    'a' 4 6
    ]

If i use sortrows(A) every column is sorted with ascending order:

A = [
    'a' 3 8
    'a' 3 16
    'a' 4 5
    'a' 4 6
    'k' 2 2
    'k' 2 4
    'k' 9 3
    'k' 9 8
    ]

Is there a way to sort rows, so first and second column is sorted with ascending order, but third one with descending? I want to get something like this:

A = [
    'a' 3 16
    'a' 3 8
    'a' 4 6
    'a' 4 5
    'k' 2 4
    'k' 2 2
    'k' 9 8
    'k' 9 3
    ]

[Cris LaPierre]

Do you want the columns sorted indepently? Or do you want to sort first by column 1, then column 2, then column 3?

[Karol Buchajczuk]

I want to sort by first column in ascending order, then second in ascending order, then third in descending, so like in example, sortrows is sorting first column, so 'a's are before the 'k's, then for every 'a' it sorts second column, so 'a' 3 is before 'a' 4 etc. My problem is sortrows sorts it by (asc,asc,asc) and I want it sorted by (asc,asc,desc).

[Star Strider]

It would likely be necessary to sort each column independently, then concatenate the results into a new matrix.

Answer 1

[Stephen23]

Where T is your table:

sortrows(T,[1,2,-3])

[Karol Buchajczuk]

Just what I wanted to do, thank you very much!

Answer 2

[David Hill]

You can't have a matrix with characters and numbers. But if you convert the characters to double, then:

A=[sort(A(:,1)),sort(A(:,2)),sort(A(:,3),'descend')];

[Karol Buchajczuk]

I'm sorry, it is actually a table, not a matrix. I've corrected my question accordingly.