How to detect differencens between 2 vectors that are sorted?

I have to following problem: I got a vector A = [1 20 3650 4000 NaN NaN] of fixed size (here 6). Given A I generate B in such a way that I generate a new value depending on the last non-NaN value of A (in this case 4000). The new Value is now 4000*rand. For example 4000*rand = 1111. So B becomes B = [1 20 1111 3650 4000 NaN], also of fixed Size 6 and sorted. More precisely the value 1111 is appended to the vector A and then sorted so that it becomes B.

But if I just have B and A given and the knwoledge of the generation process, how can I detect the 1111? My naive approch would be, looping from 1 to end through all the elements and since I know the vectors are sorted the first missmatch is the new element. But I wondered whether their is a nicer/faster solution for this?

Thanks in advance!

[Azzi Abdelmalek]

Please give a correct example, in your example B is 1x5 not 1x6 like you said, and explain how 4000*rand is located in your new matrix?

[vthuongt]

Thanks for your remark. I edited the question for clarification.

Answer 1

[Bjorn Gustavsson]

Surely this answer is what your teacher want you to learn:

setdiff(B,A)

[vthuongt]

Thank you very much. This function seems to be exactly what I needed. I have no teacher :) This problem arouse in another larger program I am working on. But anyways- how does this function performs in contrast to the for-loop? If for example I have much bigger sets?

[Bjorn Gustavsson]

I have no idea but just became curious. Please investigate and then tell us about the results.