i see your point, but in this instance I'm still thinking postgres is the way to go. this really isnt a very complex WHERE clause and I can easily verify it's returning the correct number of rows prior to running the update.

there are two ways to change the materialized datasets:
1) drop the corresponding tables in studydataset, then run that query in labkey
2) run a postgres update against that table.

If I'm already going to update exp.objectproperty directly, is there any reason not to do option 2? seems faster and less work.

Speaking as someone who manages a site, not being able to interact directly with the database is frequently a pretty big limitation. I realize that there are trade-offs and it also brings other advantages. direct postgres updates should probably be used judiciously, but they have their place. given that labkey can already interact with external schemas, i've been using that quite a bit for simple lists, rather than actually making a list. different mechanisms make sense in different places.