unique values inside a matrix

Hello,

Assuming i have a matrix like this m=

1 30

1 20

1 20

1 20

2 30

2 28

2 28

... ...

i want it to be like this:

1 30

1 20

2 30

2 28

PS: THE MATRIX IS HUGE CONSISTING OF 11000 ROWS AND 9 COLUMNS

unique is not working since it has it's taking a unique value of all, for example if i apply it for the second column it will take 28 for 1 and not for 2

thank you in advance

Answer 1

[Star Strider]

Try this:

A = [1  30
     1  20
     1  20
     1  20
     2  30
     2  28
     2  28];

Out = unique(A, 'rows', 'stable')

producing:

Out =
     1    30
     1    20
     2    30
     2    28

Experiment to get different results.

[Michel tawil]

"1 74

1 73

1 72

1 70

1 69

1 67

1 65

1 63

1 61

1 58

1 57

1 55

1 53

1 51

1 50

1 49

1 48

1 47

1 46

1 45

1 44

1 43

1 42

1 42

1 41

1 41

1 40

1 32

1 31

1 30

1 30

1 29

1 28

1 28

1 27

1 26

1 26

1 25

1 24

1 24

1 23

1 22

1 21

1 21

1 20

1 20

1 20

1 20

1 20"

after applying it, nothing changed

thanks

[Star Strider]

It changes when I run it:

A = [1	74
    1	73
    1	72
    1	70
    1	69
    1	67
    1	65
    1	63
    1	61
    1	58
    1	57
    1	55
    1	53
    1	51
    1	50
    1	49
    1	48
    1	47
    1	46
    1	45
    1	44
    1	43
    1	42
    1	42
    1	41
    1	41
    1	40
    1	32
    1	31
    1	30
    1	30
    1	29
    1	28
    1	28
    1	27
    1	26
    1	26
    1	25
    1	24
    1	24
    1	23
    1	22
    1	21
    1	21
    1	20
    1	20
    1	20
    1	20
    1	20];

Out = unique(A, 'rows', 'stable')

produces:

Out =
     1    74
     1    73
     1    72
     1    70
     1    69
     1    67
     1    65
     1    63
     1    61
     1    58
     1    57
     1    55
     1    53
     1    51
     1    50
     1    49
     1    48
     1    47
     1    46
     1    45
     1    44
     1    43
     1    42
     1    41
     1    40
     1    32
     1    31
     1    30
     1    29
     1    28
     1    27
     1    26
     1    25
     1    24
     1    23
     1    22
     1    21
     1    20

With ‘A’ being (49x2) and ‘Out’ being (38x2).

If nothing truly changes, we may not be getting all the information we need to have.

[Michel tawil]

for example the last 20 values are rounded to 20, not exactly 20, could it be a reason?

[Star Strider]

You may need to use uniquetol instead. See specifically the section on Perpare Vectors for Exact Comparison, since the 'stable' argument is not supported, and 'rows' becomes 'ByRows',1.

[Star Strider]

Using uniquetol with your .mat file:

D = load('C.mat');
A = D.C(:,1:2)
[Out,ia,ic] = uniquetol(A, 0.1, 'ByRows',1)
OutStable = A(ia,:)

produces:

A =

     1    74
     1    73
     1    72
     1    70
     1    69
     1    67
     1    65
     1    63
     1    61
     1    58
     1    57
     1    55
     1    53
     1    51
     1    50
     1    49
     1    48
     1    47
     1    46
     1    45
     1    44
     1    43
     1    42
     1    42
     1    41
     1    41
     1    40
     1    40
     1    32
     1    31
     1    30
     1    30
     1    29
     1    28
     1    28
     1    27
     1    26
     1    26
     1    25
     1    24
     1    24
     1    23
     1    22
     1    21
     1    21
     1    20
     1    20
     1    20
     1    20
     1    20
     2    85
     2    84
     2    82
     2    80
     2    78
     2    77
     2    75
     2    73
     2    72
     2    71
     2    69
     2    68
     2    67
     2    66
     2    65
     2    63
     2    62
     2    61
     2    60
     2    58
     2    57
     2    56
     2    54
     2    53
     2    52
     2    51
     2    50
     2    49
     2    48
     2    47
     2    46
     2    45
     2    44
     2    44
     2    43
     2    42
     2    42
     2    41
     2    33
     2    33
     2    33
     2    32
     3    91
     3    89
     3    87
     3    85
     3    83
     3    80
     3    78
     3    75


OutStable =
     1    20
     1    30
     1    40
     1    50
     1    61
     1    72
     2    32
     2    42
     2    52
     2    62
     2    72
     2    82
     3    75
     3    85

Experiment to get the result you want.

[Michel tawil]

yes thanks but probably the unique/uniquetol is working on the other colums not only on the first two?

[Star Strider]

Sure!

The code changes to:

D = load('C.mat');
C = D.C;
[Out,ia,ic] = uniquetol(C(:,[1 2]), 0.1, 'ByRows',1);
OutStable = C(ia,:)

producing:

OutStable =
            1           20          568         28.4            1       27.719       63.087       27.719       63.087
            1           30          755        37.75            1       27.719       63.087       27.719       63.087
            1           40          810         40.5            1       27.719       63.087       27.719       63.087
            1           50          890         44.5            1       27.719       63.087       27.719       63.087
            1           61         1005        50.25            1       27.719       63.087       27.719       63.087
            1           72         1108         55.4            1       27.719       63.087       27.719       63.087
            2           32          784         39.2            1       27.719       63.087       27.719       63.087
            2           42          831        41.55            1       27.719       63.087       27.719       63.087
            2           52          904         45.2            1       27.719       63.087       27.719       63.087
            2           62          983        49.15            1       27.719       63.087       27.719       63.087
            2           72         1020           51            1       27.719       63.087       27.719       63.087
            2           82         1068         53.4            1       27.719       63.087       27.719       63.087
            3           75         1040           52            1       27.719       63.087       27.719       63.088
            3           85         1106         55.3            1       27.719       63.087       27.719       63.088

This is the shortg format option, since it’s easier to read (and post). The full precision remains internally.

[Michel tawil]

Thank you!

[Star Strider]

As always, my pleasure!

Answer 2

[Guillaume]

I'm not sure what you mean by unique is not working. You haven't explained the reasoning for going from your input to your output, but it does look like you want unique rows:

result = unique(m, 'rows')

With your demo data:

>> m = [
1  30
1  20
1  20 
1  20
2  30
2  28
2  28];
>> unique(m, 'rows')
ans =
     1    20
     1    30
     2    28
     2    30
>> unique(m, 'rows', 'stable')  %with stable to get the same order as your example
ans =
     1    30
     1    20
     2    30
     2    28

[Michel tawil]

I applied it, it looks the same as mr.star's answer, for some reason it didn't work.

if u look in the end of the answer i gave to him, u can see that I have 5 of 20s but they weren't orignally 20s, I applied the command 'round' before ( round(19.8)->20), could it be a reason of not detecting it ?

[Guillaume]

The simplest way for us to understand what you're talking about is to attach a mat file with your data to your question.

>> unique([20; round(19.8)])
ans =
    20

As you can see it's not a problem unique returns just one value. Whatever your data is, it's not exactly how you describe it.

Possibly, it's something like:

>> v = [20; 20+1e-14]
v =
                        20
                        20
>> unique(v)  %return 2 elements because the 2nd one is not exactly 20.
ans =
                        20
                        20
>> ans - 20  
ans =
                         0
      1.06581410364015e-14                        

[Michel tawil]

hello, please find it attached here, i noticed that other columns could have non-unique values but i want the unique function to be just applied for all the rows but for columns 1 and 2 ? idk if this is possible

[Michel tawil]

i can use this and it works well but i lose the cotents of the other colums

M=unique(C(:,1:2),'rows','stable');

[Michel tawil]

Thank you !