Hi, I have an SD model which has a bunch of dimensioned inputs which values change over time. The changes however are not triggered by the model's logic - they are known (values on a per time basis for all subscripts are stored in a database) before the model starts executing. I am planning to read in these inputs from the database before the model starts running and load them in memory. My question is whether I can prepopulate the SD model inputs for all time periods before the model starts or I have to populate on per time period basis as the model runs? My guess is that I cannot do the latter, and have to settle for the former. If that's the case, I most likely will load the values in a data structure that allows for fast lookups on a per input per time basis. The question is how do I catch the event of the model progressing from one period to the next? Thanks, -nik