you can't technically have multiple schemas in a single instance. you can partition your schema according to some prefix so that you have independent subunits, but you still have to write the entire thing all at once. if you want multiple schemas you need to run multiple clusters.