# Remove array values with multiple occurances from "parent" array

2 views (last 30 days)
Joe I on 9 Jul 2020
Edited: Bruno Luong on 10 Jul 2020
How do you remove values in an array from a "master database" array? For example:
a = [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4]
b = [1, 1, 2, 4, 4, 4]
I would like to remove each occurance of each value in 'b' from 'a' to get:
c = 1, 2, 2, 3, 3, 3
setdiff(a, b) doesn't work because it removes all instances from 'a' including the repeating values (it even says no repetitions in the documentation):
c = 3
[~, col] = ismember(b, a)
a(col) = []
doesn't work because it only removes one instance of the numbers (the first index where it occured):
c = 1, 1, 2, 2, 3, 3, 3, 4, 4
Thanks for the help.

David Hill on 9 Jul 2020
Edited: David Hill on 9 Jul 2020
a = [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4];
b = [1, 1, 2, 4, 4, 4];
c = a;
for i=b
c(find(c==i,1))=[];
end
Joe I on 9 Jul 2020
Awesome, thanks! Works great, scales up to my datasets with tens of thousands of values in each. Seems so obvious now after seeing it. Just be sure that each variable is a row vector.

Bruno Luong on 10 Jul 2020
Edited: Bruno Luong on 10 Jul 2020
Test samples
a = [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4]
b = [1, 1, 2, 4, 4, 4]
Code
u = unique([a,b]);
n = length(u);
[~, ia] = ismember(a,u); na = accumarray(ia(:),1,[n 1]);
[~, ib] = ismember(b,u); nb = accumarray(ib(:),1,[n 1]);
c = repelem(u,max(na-nb,0))
Result
c =
1 2 2 3 3 3