Didn’t realize you were expecting possibly mulitple matches.

–Loren

Could you elaborate on how you think unique could help?

Exact equality for floating point numbers is very iffy. You will need to build in a tolerance test.

Could you use unique looking for rows somehow with 3 outputs A and B?

–Loren

I have A= 80,000×2 i.e.

[275.0919,-13.6656;275.1045,-13.8508...... ]

and B= 10,000×2 i.e.

[275.1045,-13.8508;.... ]

and need to find every occurance where the value in B occurs in A prefereable with the index for both included. My task is then to compare other values relating to these indexes in other arrays. These are actually coordinates and I need the ones that match in both, but there could always be a slight difference maybe the last decimal that would still need to be accepted. This sounds simple to do in my head and I am not a MATLAB expert by any means so any help in simple English would help. Thanks.

Welcome to the world of floating point arithmetic. The numbers you are printing out are not the full picture. If you use format long, you would likely see differences between the first 2 entities in your vector. == is only true if the numbers are non NaN, and are bitwise exactly equal. It’s much better to allow a tolerance when comparing floating point numbers. If you look through the blog for items with category numerical accuracy, you should see several relevant posts.

–Loren

unique does not stop repition of 0.2000 in this case.But in another similar case of 0.2000, it stops repetition.What is the solution

Why so complicated? You can just do

[row, col] = find(A==value)

no ismember needed…

–Loren

A = [11 22 34 56 89
23 44 11 20 66
79 54 32 17 89
11 66 21 45 90];

value = 22; %value to be searched

[row,col]=find(ismember(A,[value]))