How to use knn classification algorithm to find the nearest neighbour

we have a excel sheet which is our dataset and another excel sheet(say training set) having few but same values from the original data set. and now we want to train our system to be able to classify that whether a value from the training set belongs to the original set

[Akira Agata]

If my understanding is correct, your dataset, say dataD, is N1-by-m matrix and training set, say dataT, is N2-by-m matrix with N2 < N1. And some rows of dataT, for example 2nd and 4th, are equal to corresponding rows of dataD, for example 100th and 200th.

So, dataT(2,:) = dataD(100,:) and dataT(4,:) = dataD(200,:).

And what you want to do is to find these two rows. Is this my understanding correct??