I have a problem that I need to calculate Kullback–Leibler divergence, but the problem is that I have some elements that are $0$. Is there a way that I'm able to deal with situations like that? I know that Wiki is saying that if $P(i)=0$, then $Q(i)$ has to be $0$ as well, but maybe there is an accepted workaround to get rid of the situation.
For example, I have two arrays $[1,2,3,0,6,7,0]$ and $[0,2,3,5,0,0,9]$.