Can you optimize this row matching code?

Hi everyone,

I have a huge matrix on which I need to do some matching operation. Here's the code I have written and it works fine, but I think there is some room to make it more optimized or write another code that does the matching in less time. Could you please help me with that?

rowsMatched = find(bigMatrix(:, 1) == matchingRow(1, 1) & bigMatrix(:, 2) == matchingRow(1, 2) & bigMatrix(:, 3) == matchingRow(1, 3))

The problem with this code is that I cannot use the && operand. So, in case one of the columns do not match the program still checks the next condition. How do you think I can avoid this?

Thank you

[Matt Kindig]

Is matchingRow only a 1x3 vector? Or do you loop through all rows in matchingRow?

Can you post your full code?

[Amir]

It's just a 1x3 vector in this case, but it could be up to 1 x 10 (in case I had a bigMatrix with the size, say, (1m, 10). However, I still don't know how to make the code that much flexible being able to do the matching for a 1 x 3 up to 1 x 10 matchingRow vector.

Answer 1

[Amir]

Here's the solution:

rowsMatched = find(all(bsxfun(@eq, bigMatrix, matchingRow),2))

http://stackoverflow.com/questions/23252389/optimize-this-row-matching-code

Answer 2

[Matt Kindig]

How about this approach? Is it fast enough?

%bigMatrix is MxN, matchingRow is 1xN
rowsMatched = find(ismember( bigMatrix(:,1:3), matchingRow, 'rows'));

[Amir]

No this is slower.