how to compare column values of a matrix with a column vector of different size?

Hello everyone, how can I compare a 10x5 matrix with a row vector, let me explain it better.

I have the vector

detected = [56

40

33

31

28

13

10

1]

and the matrix

M_transmision = 1 40 -84 -638 644

2 1 138 -276 308

3 31 -74 -157 173

4 28 -274 -511 580

5 13 53 -35 64

6 13 124 -367 388

7 56 30 -290 292

8 33 20 -263 263

9 28 -504 103 515

10 10 -118 -226 255

I need to compare the unique values of the detected vector with column 2 of the M_transmission matrix, and extract column 5 of the matrix from each unique value found.

for example the result would be the following.

result = [56 292

40 644

33 263

31 173

28 515

13 64

10 255

1 308]

note: If the numbers are repeated, as shown in this example, the number 13 and 28 are repeated, you have to extract the smallest number from column 5 of the matrix.

for instance

number 13 has:

13 64

13,388

and the number 28 has:

28,580

28,515

has to extract the smallest value, that is:

13 64 and 28 515

any help i will appreciate it

I was trying to do it like this, but I can't

detected(:,end+1:5)=missing;
detected = array2table(detected);
M_transmision = array2table(M_transmision);
result2 =innerjoin(M_transmision,detected,'LeftKeys',["M_transmision2"],'RightKeys',["detected1"]);

Answer 1

[Adam Danz]

Load data

detected =   [56
    40
    33
    31
    28
    13
    10
    1];

M_transmision = [1	 40	 -84    -638	   644
    2	 1	 138    -276   308
    3	 31	 -74    -157   173
    4	 28	 -274   -511   580
    5	 13	 53	    -35	   64
    6	 13	 124    -367   388
    7	 56	 30	   -290	   292
     8	 33	 20	   -263	   263
    9	 28	-504    103	   515
    10  10	-118   -226	   255]; %   

Return minimum of column 5 within groups defined by detected and column 2.

[~, idx] = ismember(M_transmision(:,2),detected);
y = splitapply(@min,M_transmision(:,5),idx); 
result = [detected(unique(idx)), y]

result = 8×2

56 292 40 644 33 263 31 173 28 515 13 64 10 255 1 308

However, if there is a value from detected that is not in column 2 of M_transision, then this will cause an error. I assume detected = unique(M_transmision(:,2)) in which case, this won't be a problem.

Update

This version deals with mismatches between the two vectors better.

[ism, midx] = ismember(M_transmision(:,2),detected);
idx = findgroups(midx(ism));
y = splitapply(@min,M_transmision(ism,5),idx); 
result = [detected(unique(idx)), y]

result = 8×2

56 292 40 644 33 263 31 173 28 515 13 64 10 255 1 308

[giancarlo maldonado cardenas]

hello @Adam Danz, according to what I understand this line y = splitapply(@min,M_transmision(:,5),idx); calculate the minimum of the repeated numbers right?

How can I do in the event that there are no repeated numbers, that is, I need both ways, my numbers are random, sometimes they are repeated, sometimes they are not.

you can help?

[Adam Danz]

The splitapply line returns the minimum value in colum 5 for rows that have the same value in column 2. If there are no repeat values in column 2, then it just returns the value in column 5.

For example,

min([4,3])

ans = 3

min(3)

ans = 3

If this doesn't address your follow-up questions, please provide an example where my solution fails.

[giancarlo maldonado cardenas]

in this case, for example, the detected vector and M_transmission, are always random, and sometimes there is an error in that line.

attached new vector detected.mat and M_transmision.mat

detected = detected.'; to make column vector.

[giancarlo maldonado cardenas]

detected = [61

49

40

34

30

26

25

18

8]

M_transmision = [1 61 -301 36 304

2 30 46 -50 68

3 26 -28 -189 191

4 49 -136 307 336

5 42 255 271 372

6 25 -167 -39 171

7 8 269 -306 407

8 34 -314 59 319

9 18 -49 107 118

10 40 -77 75 108]

In this case, an error also appears.

Error using splitapply (line 61)

Group numbers must be a vector of positive integers, and cannot be a sparse

vector.

[Adam Danz]

Try the updated answer and let me know if the problem persists.

[giancarlo maldonado cardenas]

detected = [63

60

52

43

27

24

18

15

6

4]

M_ transmision = [ 1 6 -438 -218 489

2 63 34 4 34

3 4 120 -49 129

4 5 51 -57 77

5 43 -343 -257 429

6 18 -469 331 574

7 60 -503 16 503

8 24 -97 31 102

9 15 346 -148 376

10 52 -95 235 253

in this case a size error appears, in this line idx = findgroups(midx(ism));

cut my vector from 1x10 to 1x9, I don't know why!

my detected vector should not be cut, it should only be compared if the values that were detected exist in the transmission matrix and extract their distance. then I have to compare the distances in another process.

but there is missing a detected value.

[Adam Danz]

> in this case a size error appears, in this line idx = findgroups(midx(ism))

I copied your inputs and added my solution and ran the results below. I don't see an error message.

> cut my vector from 1x10 to 1x9, I don't know why!

Hint: detected does not contain 5.

detected = [63
    60
    52
    43
    27
    24
    18
    15
    6
    4];

M_transmision = [           1	6     -438	-218	  489
    2	63	34	4	  34
    3	4	120	-49	 129
    4	5	51	-57	 77
    5	43	-343	-257	 429
    6	18	-469	331	 574
    7	60	-503	16	 503
    8	24	-97	31	 102
    9	15	346	-148	 376
    10	52	-95	235	 253];

[ism, midx] = ismember(M_transmision(:,2),detected);
idx = findgroups(midx(ism));
y = splitapply(@min,M_transmision(ism,5),idx); 
result = [detected(unique(idx)), y]

result = 9×2

63 34 60 503 52 253 43 429 27 102 24 574 18 376 15 489 6 129

[giancarlo maldonado cardenas]

Excuse me, I think I did not express myself well, at first the detected vector has 10 values, then in the line line idx = findgroups(midx(ism)). Cut my detected vector from 10 values to 9, would I have to continue having the 10 values, because I cannot eliminate any detected value from the detected vector, do I understand?

actually it would not be an error, it simply cannot eliminate a detected value, because the system already knows that there are 10 values, later I have to compare the distance of the X detected values, in this case the 10 values.

[Adam Danz]

Column 2 of the M_transmision matrix contains a 5 but the detected vector does not contain a 5. Where should [5,77] go in the results?

[giancarlo maldonado cardenas]

thank you very much, I had to modify some things, to make it work in the following process.