Match row vector in matlab efficiently

[Mindscrape]

I would like to take a row vector of my choosing and see if it matches any other rows in say a 1e6 x 100 array. I want to do this as fast as possible with matlab. There should only be, if any, one match with the actual data I'll be using because every row in my data (the matrix) will be unique.

What I've come up so far is to simulate the unique rows with rand, and the fastest method seems to using bsxfun (I'm getting ~.3 seconds on my machine).

A=ceil(10*rand(1e6,100));
B=ceil(10*rand(1,100));
tic
index=find(all(bsxfun(@eq,A,B),2));
toc
index

this is faster than both using ismember and using a for loop to check each row one-by-one. Still, there is something that is bugging me, something that is telling me that there should be an even more elegant and quicker way to find if there is a match and where that match is.

 

[Valkarie]

If anyone has any suggestions, I'd be more than happy to hear them.The best way to do this would be to use a hashing algorithm. A hashing algorithm takes the data, such as a row vector in your case, and converts it into a numerical value that can then be used to quickly compare and find matches. These algorithms are designed to be very fast and efficient, so they should be able to give you much better results than using bsxfun or any other methods. There are many different hashing algorithms available, but some of the most popular ones are MD5, SHA-1, and SHA-2.