
It takes ~1.5 hours to read the 375m active fire files over the network (in the case above, to produce a shapefile). Since this information is the same for all the runs, it makes more sense to store the active fire table once, rather than once per schema.