Deleting rows with identical values in determined columns of two matrices.

Hello. I have two large matrices with the next dimensions

Matrix a = 6700 x 8.

Matrix b= 388000 x 8

As a note, I would like to say that both "matrix a" and "matrix b", column 1 to column 4 are time data (year, month, day, hour), and from column 5 to column 8 are data measurements.

It is required a routine which compares column 1 to column 4 from "matrix a" with column 1 to column 4 from "matrix b", finds the elements repeated in column 1, column 2, column 3 and column 4 of both matrices, and finally deletes all this rows only in "matrix b", generating a "new matrix b". Also is desirable that the "new matrix b" has an ascending order taking into account only column 1 to column 4, due to this columns are the time data.

For instance:

Matrix a = [2009,10,2,2,0,0,0,0; 2009,10,2,2,0,0,0,0; 2009,10,9,5,0,0,0,0]

Matrix b = [2012,10,9,5,1,2,3,4; 2009,10,9,5,3,9,7,1; 2012,10,9,5,1,2,3,4]

New Matrix b => [2012,10,9,5,1,2,3,4; 2012,10,9,5,1,2,3,4]

In this example, "matrix b" has in the second row, the same elements that are present in the third row of "matrix a", only taking into account column 1 to column 4 of both matrices (the data repeated are 2009,10,9,5). For this reason, the entire second row of "matrix b" is eliminated in the "new matrix b".

I want to implement this kind of solution to my whole "matrix b", which has dimensions of 388000 x 8.

Thanks in advance.

Answer 1

[Azzi Abdelmalek]

a = [2009,10,2,2,0,0,0,0; 2009,10,2,2,0,0,0,0; 2009,10,9,5,0,0,0,0]
b = [2012,10,9,5,1,2,3,4; 2009,10,9,5,3,9,7,1; 2012,10,9,5,1,2,3,4]
out=b(~ismember(b(:,1:4),a(:,1:4),'rows'),:)

[Miguel L]

Thaks again Azzi it works pretty well.