Right, I think the timestamp with timezone type is a great example of one of these short comings. I’ve generally used it as a step in a data pipeline, where so

Right, I think the timestamp with timezone type is a great example of one of these short comings. I’ve generally used it as a step in a data pipeline, where some intermittent state or aggregation where being stale/stable for a period of time is ok. Did this recently for a daily trends report on some news data - live querying the data is slow, but baking it off into a MV makes it much faster and keeps the query result from changing too rapidly to new data.