Consider a quantity $[v_1,v_2,\dots,v_N]$ where each component is a function $$v_1=v_1(x_1,x_2,x_3,x_4),v_2=v_2(x_1,x_2,x_3,x_4),\dots,v_N=v_N(x_1,x_2,x_3,x_4).$$ Assume that a vector of discrete values $[x]_1,$ $[x]_2,$ $[x]_3,$ $[x]_4$ is given for each independent variable and the function values are known at the points $$v_n^{ijkl}=v_n(x_1^i,x_2^j,x_3^k,x_4^l).$$ The function values $v_n^{ijkl}$ can be assembled into vectors $$[v]_n\equiv[v_n^{1111},v_n^{1112},v_n^{1113},\dots,v_n^{1121},v_n^{1122},v_n^{1123},\dots ].$$
First, the values of the independent variables $[x]_\mathtt{IV}$ are entered by batches
IV JIV T IV V $[x]_\mathtt{IV}$
where
IV batch number,The quantity $[v_1,v_2,\dots,v_N]$ is entered by the batch
XY ISET T KQT I [JIV] V $[v]_1$ $\dots$ V $[v]_N$
where
XY are the set key letters,IV batch numbers specifying the kind and order of variables $x_1,x_2,x_3,x_4.$
The dependent variable for an argument outside the range specified by the IV batch is replaced by its value for the closest defined argument (extrapolation by constant). A similar rule applies in the case of dependence on multiple arguments.