Finding elements of matrix that match a condition

Hello everyone,

I am working with some big data sets and I wanted to find some elements from one matrix that satisfy a condition. So, what I have. Matrix A is 251576x5. Column 1 is a timestamp, columns 2 to 4 are x,y,z coordinates and column 5 is something else that is not important to me. Matrix B is 26300x1 and is a column of timestamps starting from 0 and incrementing by 100. What I want is, for each timestamp on B I want the row with that same, or the closest possible, timestamp from A saved in another Matrix. So my output should be a matrix C, 26300x5 in dimensions.

For example, this is a piece of matrix A

25194.41 313.93 440.32 372.93 274

25727.49 327.04 420.64 387.72 277

A timestamp from matrix B is

25200

So what would be saved here in matrix C should be 25194.41 313.93 440.32 372.93 274.

I hope I made myself clear enough. Thank you all in advance

Answer 1

[Ameer Hamza]

You can use interp1() with nearest option

Bx = interp1(A(:,1), A(:,2:4), B, 'nearest');
C = [B Bx];

[Ioannis Kiratzis]

Thank you for your swift reply. I am getting this error message

Error using griddedInterpolant

The grid vectors must contain unique points.

Error in interp1 (line 148)

F = griddedInterpolant(Xext,V,method);

Error in my_script (line 31)

Bx = interp1(A(:,1), A(:,2:4), B, 'nearest');

[Ameer Hamza]

Are there duplicate timestamps in first column of matrix A? Can you share your matrices in a .mat file?

[Ioannis Kiratzis]

It appears to work only if A and B have the same number of rows

[Ameer Hamza]

No, it works when they have different number of rows too

A = [linspace(0, 100).' rand(100, 3)];
B = linspace(0, 100, 1000).';

Bx = interp1(A(:,1), A(:,2:4), B, 'nearest');
C = [B Bx];

[Ioannis Kiratzis]

I had to chop matrix A down a bit but it shouldn't make a difference. Choped due to limitations of upload file size

[Ioannis Kiratzis]

It works even for different sizes you are right. It's weird. it works when I chop A down to 100000 lines but if I go above it doesn't work anymore

[Ameer Hamza]

The code works correct on the matrices you shared? Can you share matrices where it fails? Also check of there are any duplicate time stamps in matrix A.

[Ioannis Kiratzis]

I can't share thethe complete matrices were it fails as they exceed the limit on the file size I can upload. I can work around the problem. I could work on a smaller portion of the dataset anyway since these are results from long experiments. So you are saying the problem is the presence of duplicate timestamps in A?

[Ameer Hamza]

Yes. If that happens, then the interpolation will fail. I guess that is the most likely reason for the error you are receiving.