Currently, I believe that we are frowning on writing things directly to Public by default.
Also, we had already taken that step in our systems. Furthermore we force Public to be the first Schema, so this restriction enforces that everything created must be assigned to the appropriate schema.
That could make us a bit more sensitive to how pgbench operates. But this becomes an opportunity to sharpen a tool.
But logging in to our regular users it has no ability to create it's tables and do it's work. At the same time, we want exactly this login, because we want to benchmark things in our application.
Our lives would be simplified if there was a simple way to specify a schema for pgbench to use. It would always reference that schema, and it would not be in the path. Making it operate "just like our apps" while benching marking our items.
While we are here, SHOULD we consider having pgbench have a default schema name it creates and then cleans up? And then if someone wants to override that, they can?
Kirk...
