find lagged values of variables

Hi,

I have a matrix of data in which and observation is defined by i (people), j (alternatives), t (time). For each i, j, and t there is a random variable u_{ijt}, and I would like to generate a vector with lagged values of u. So, I have

i=[1 1 1 1 1 1 1 1 2 2 2 2 2]';
j=[1 2 3 1 2 1 2 4 1 2 1 2 4]';
t=[1 1 1 2 2 3 3 3 2 2 3 3 3]';

And I have u=[u_{111} u_{121} u_{131} u_{112} u_{112} u_{122} u_{113} u_{123} u_{143} u_{212} u_{222} u_{213} u_{223} u_{243}]'; where the u_{}'s are, of course, numbers.

Now I want to define the lagged value of u for every i and j, over t. In case the lagged value doesn't exist, as it is the case for j=4 in period 3, then replace by zero.

How can I do this?

Thanks