Duplicated rows make use sortdifficult, but in this case you can rely on what uniqueworks for arrays of row cells and both types of output and optionally returns the indices of these sorted elements in the original input:
>> a = {'abc' 'aty' 'utf8' 'sport' 'utf8' 'abc'}
a =
{
[1,1] = abc
[1,2] = aty
[1,3] = utf8
[1,4] = sport
[1,5] = utf8
[1,6] = abc
}
>> [b, ~, index] = unique(a)
b =
{
[1,1] = abc
[1,2] = aty
[1,3] = sport
[1,4] = utf8
}
index =
1 2 4 3 4 1
[~, ~, index] = unique(a);, .