# how to remove repeating rows from a matrix?(there is a case)

1 view (last 30 days)
Zhang on 13 Oct 2011
I want to combine all the possible no-repeating permutations of two or more 0,1 row vectors, for example, there are a=[1,0], b=[1,1,0], the combination of all their possible no-repeating permutations will be
0 1 0 1 1
0 1 1 0 1
0 1 1 1 0
1 0 0 1 1
1 0 1 0 1
1 0 1 1 0
I use the following script to ahieve this,
a=[1,0];
b=[1,1,0];
A=perms(a); %all the possible permutations, may contain repeating rows
B=perms(b);
[r_a,c_a]=size(A);
[r_b,c_b]=size(B);
AB=[];
for i=1:r_a
for j=1:r_b
COM1=[A(i,:) B(j,:)]; %combine ith rows in A with all rows in B
D1(j,1:c_a+c_b)=COM1;
end
AB=cat(1,AB,D1); %update AB on every ith loop
end
AB=unique(AB,'rows'); %delete repeating rows
% AB=union(AB,AB,'rows'); %union can also be used to delete repeating rows
but,in my real case, there are
a=[1,1,0,0,0];
b=[1,1,1,1,0,0];
c=[1,1,1,1,0,0,0];
d=[1,1,1,0,0];
e=[1,1,0,0];
I want to combine all the the possible no-repeating permutations(the result will be ABCDE), using method like the above script,
A=perms(a);
B=perms(b);
C=perms(c);
D=perms(d);
E=perms(e);
and then,
Step 1, combine A and B, delete repeating rows, and result in AB;
Step 2, combine AB and C, delete repeating rows, and result in ABC;
Step 3, combine ABC and D, delete repeating rows, and result in ABCD;
Step 4, combine ABCD and E, delete repeating rows, and result in ABCDE;
but using my method(either using "union" or "unique" to delete), when it comes to Step 4, it is out of memory,
Elapsed time is 1.452821 seconds.%Step 1 Elapsed time is 19.556511 seconds.%Step 2 Elapsed time is 456.368700 seconds.%Step 3 ??? Out of memory. Type HELP MEMORY for your options.
Error in ==> unique at 201 d = b(1:rows-1,:)~=b(2:rows,:);
Error in ==> combine_perms_5 at 85 ABCDE=unique(ABCDE,'rows');%here, I use unique
my questions are,
Q1 Are there simple ways to combine all the possible no-repeating permutations of a,b,c,d,e?
Q2 Are there simple ways to delete repeating rows, making sure that it is not out of memory?
I really appreciate your comment on this!

Andrei Bobrov on 13 Oct 2011
A = perms([0 0 1 1 1]);
AB=unique(A(sum(A(:,1:2),2)==1 | sum(A(:,3:end),2)==2,:),'rows');
variant
a=[1,1,0,0,0];
b=[1,1,1,1,0,0];
c=[1,1,1,1,0,0,0];
d=[1,1,1,0,0];
e=[1,1,0,0];
Din = {a b c d e};
d = cellfun(@(x)unique(perms(x),'rows'),Din,'un',0);
ns = cellfun('size',d,1);
inarg = arrayfun(@(i1)1:i1,ns,'un',0);
M = cell(1,numel(Din));
[M{:}] = ndgrid(inarg{:});
outc = cellfun(@(x,y)x(y(:),:),d,M,'un',0);
out = [outc{:}];
solution on my computer (HP Compaq dx2400 Microtower IntelC2D 2.13GHz, RAM 2GB)
##### 1 CommentShowHide None
Zhang on 13 Oct 2011
It works, thx indeed! another question, I run the script and the result variable "out" in workspace is an 315000*27 double variable, but "Min"and "Max" shows"Too many elements", I run max(out) and min(out), it shows,
>> max(out)
ans =
Columns 1 through 11
1 1 1 1 1 1 1 1 1 1 1
Columns 12 through 22
1 1 1 1 1 1 1 1 1 1 1
Columns 23 through 27
1 1 1 1 1
>> min(out)
ans =
Columns 1 through 11
0 0 0 0 0 0 0 0 0 0 0
Columns 12 through 22
0 0 0 0 0 0 0 0 0 0 0
Columns 23 through 27
0 0 0 0 0
how can it be all 1 or all 0 in "out"?