Wrong output of ismember command

Dear all,

In part of my code I am using ismember command and as the array I am using has 21*9 shape and all of these elements are in the first element I expected an array of 21*9 members to be my out put while the output has shape of 22*9! Any idea about the cause? Please find my code in the following. I found the rows having same value in the 9th column in all_com variable and then used idx commands to find the rows corresponding to that.

Thank you in advance.

in1=load('in1.mat')

in1 = struct with fields:
    in1: [265×9 double]

in2=load('in2.mat')

in2 = struct with fields:
    in2: [251×9 double]

out1=load('out1.mat')

out1 = struct with fields:
    out1: [271×9 double]

out2=load('out2.mat')

out2 = struct with fields:
    out2: [256×9 double]

all_common= intersect(intersect(intersect(in1(:,9), in2(:,9)), out1(:,9)), out2(:,9));

Index in position 2 exceeds array bounds. Index must not exceed 1.

idx1 = ismember(in1(:,9),all_common);
in1_com = in1(idx1,:);

idx2 = ismember(in2(:,9),all_common);
in2_com = in2(idx2,:);

idx3 = ismember(out1(:,9),all_common);
out1_com = out1(idx3,:);

idx4 = ismember(out2(:,9),all_common);
out2_com = out2(idx4,:);

Answer 1

[Walter Roberson]

filedata = load('in1.mat');
in1 = filedata.in1;

filedata = load('in2.mat');
in2 = filedata.in2;

filedata = load('out1.mat');
out1 = filedata.out1;

filedata = load('out2.mat');
out2 = filedata.out2;

all_common= intersect(intersect(intersect(in1(:,9), in2(:,9)), out1(:,9)), out2(:,9));
idx1 = ismember(in1(:,9),all_common);
in1_com = in1(idx1,:);

idx2 = ismember(in2(:,9),all_common);
in2_com = in2(idx2,:);

idx3 = ismember(out1(:,9),all_common);
out1_com = out1(idx3,:);

idx4 = ismember(out2(:,9),all_common);
out2_com = out2(idx4,:);

[Hamid]

Thank you for the comment but the problem is not solved. out2_com has a shape of 22*9 while it should be 21*9.

[Walter Roberson]

out2(69,9) and out2(235,9) are both 13740001

You have 256 entries in out2(:,9) but only 247 of them are unique

[Hamid]

@Walter RobersonThank you for your reply.

As all_common has 21*1 I expected all (in1_com, in2_com, out1_com and out2_com) to have 21*9 shape while for out2_com the shape is 22*9.

[Walter Roberson]

out2(69,9) and out2(235,9) are both 13740001 which is a member of all_common. In order to get the 21 results that you expect, you would need to decide between selecting row 69 or row 235; how do you want to decide which of the two to use?

Your logic of expecting the same number of rows as entries in all_common assumes that the entries in column 9 are unique, but they are not.

[Hamid]

@Walter RobersonThanks a lot. I expected the elements in column 9 to be unique according to my results but it was no correct.