How to delete a row if it doesn't follow a pattern?

Hi, I have... a = [ 2 5 1; 3 6 2; 3 4 1; 9 4 2; 8 3 1; 3 2 2; 9 5 2; 4 8 1]

Notice how the last column follows a pattern of 1, 2,1,2..and so on. The 6th & 7th row both have a 2 in the last column...thus does not follow the pattern. How would I delete the 6th row and any other row if it does not follow the 1,2,1,2 pattern of the last column?

*If there are repeated numbers in the last column I would want to keep the last row with the repeated number and delete the others Thanks.

Answer 1

[Azzi Abdelmalek]

You can adapt the answer of your previous question http://www.mathworks.com/matlabcentral/answers/229434-delete-row-not-following-pattern#comment_297496

id=[1 diff(a(:,3))' ]
ii=find(id==0)
for k=1:numel(ii)
  idx(k)=ii(k)-1
end
a(idx,:)=[]

[Sha S]

When I tried this I get: "Undefined function or variable 'idx'."

[Brendan Hamm]

That must mean there are no places which satisfy:

id == 0

When this is the case the ii will be empty, the loop never entered and finally idx will not be initialized:

id = [1 2 5];
ii = find(id == 0)
ii = 
  []
numel(ii)
ans =
   0

So just initialize idx outside of the loop to the empty array

id=[1 diff(a(:,3))' ]
ii=find(id==0)
idx = []; % ADD THIS LINE
for k=1:numel(ii)
  idx(k)=ii(k)-1
end
a(idx,:)=[]