What Is Vector Trap?

What is the most efficient way to test whether a matrix contains a given row vector in MatLab?

• More specifically, that a given 1 by n row vector exists as a row in an m by n matrix, such that all elements are in the same order and are equal. The ismember() function is very slow, and my own attempt is better but still quite slow: function index = findIndexOfVectorInMatrix( matrix, vector ) % Finds the index of 'vector' as it appears in 'matrix', such that % >> index = findIndexOfVectorInMatrix( matrix, vector ); % >> matrix( index(1), : ) % returns the original 'vector'. % 'index' may itself be a vector of indices if there are % multiple copies of 'vector' in 'matrix'. [r,c]=size(vector); if r ~= 1 && c ~= 1 error('Vector must have one dimension of size = 1.'); end functionalString = 'find( sum( '; if r == 1 % Rows of matrix are vectors. functionalString = strcat( functionalString, [' ( matrix( :, 1 ) == vector( 1 ) )'] ); for ii = 2:c functionalString = strcat( functionalString, ... [' + ( matrix( :, ', num2str(ii),' ) == vector( ', num2str(ii),' ) )'] ); end functionalString = strcat( functionalString, ', 2 ) == c );' ); elseif c == 1 % Columns of matrix are vectors. functionalString = strcat( functionalString, [' ( matrix( 1, : ) == vector( 1 ) )'] ); for ii = 2:r functionalString = strcat( functionalString, ... [' + ( matrix( ', num2str(ii),', : ) == vector( ', num2str(ii),' ) )'] ); end functionalString = strcat( functionalString, ', 1 ) == r );' ); end [index] = eval( functionalString );

• Answer:

  First post so I'm not sure if this will show up with the correct formatting but how about something like getting the absolute value of subtracting repmat(row vector,n,1) from your matrix, then getting the norm of each row vector and reporting if any norm =0. I tested the elapsed times of calculation for a magic 2000x2000 matrix searching for the last row using the following methods: 1. [~,index]=ismember(matrix, vector, 'rows') Elapsed time is 2.892619 seconds. 2. for i=1:numrows if ismember(matrix(i,:),vector) disp(i) end end Elapsed time is 1.060776 seconds. 3. absdiff=abs(matrix-repmat(row, numrows,1)); for i=1:numrows if norm(absdiff(i,:))==0 disp(i) end end Elapsed time is 0.363086 seconds.