remove duplicate rows from a matrix

I have a matrix of the form

A=[317.0000  282.0000  310.0000  259.0000  257.0000  305.0000  294.6667  282.0000
   317.0000  282.0000  309.0000  372.0000  257.0000  305.0000  294.3333  319.6667
   317.0000  282.0000  257.0000  305.0000  310.0000  259.0000  294.6667  282.0000
   317.0000  282.0000  257.0000  305.0000  309.0000  372.0000  294.3333  319.6667
   92.0000  166.0000   55.0000  235.0000   71.0000  173.0000   72.6667  191.3333
   92.0000  166.0000   71.0000  173.0000   55.0000  235.0000   72.6667  191.3333];

I want to remove the redundant rows from A. Can anyone help. TIA.

Answer 1

[Thorsten]

unique(sort(A,2), 'rows')

[Ananya Malik]

Thnaks a lot @Thorsten. Works perfectly.

[John Allen]

warning, if position in the row is important then you dont want this. ie if (1,2) is different to (2,1)

Answer 2

[Walter Roberson]

uA = unique(A, 'rows', 'stable');

[Ananya Malik]

Does not work. It returns back exactly A.

[Guillaume]

"Does not work"

It does exactly what you asked. If it does not work that would be because no two rows are exactly identical. There may be some small differences between the values that matlab does not show by default.

You can use uniquetol instead of unique to make up for these small differences:

uA = uniquetol(A, 'ByRows', true); %use default tolerance. You can specify your own

There is no 'stable' option for uniquetol. The returned rows will be ordered.

[Ananya Malik]

A simpler example

A = [1 2 3 45; 3 2 1 45; 1 4 5 54; 5 4 1 54];

I need the output

   B= [1 2 3 45; 1 4 5 54]

[Ananya Malik]

Thanks Guillaume for your input.

[Walter Roberson]

You have a row that begins with 5, but no output row that begins with 5. You have removed too much.

[John Allen]

the post title is misleading - you don't want to remove duplicate rows, you want to remove rows with the same contents, irrespective of order

[Walter Roberson]

[~, iA] = uniquetol(A, 'byrows', true);
A(sort(iA), :)      %the unique rows, in the original order