Actually, you can also include t and keep the original x2 data, treat x2 as factor as well. The following are the code I referenced floor 4, but give a little change. You can try both. choose the one you like.
Good Luck.
data aaa;
input x1 x2 x3 t y;
datalines;
2.5 0 0 3.4 0
1.2 2 0 3.73 0
173 2 0 3.73 0
3.5 0 0 3.83 0
119 2 0 4 0
39.7 0 0 4.03 0
... ... ... ...
;
run;
proc logistic data=aaa Descending;
class x2 x3/PARAM=REFERENCE Descending;
model y=x1 x2 x3 t;
run;