The casing of names seems to be part of the schema factory rather than the convention.
Schema factory is the ‘front door’ of an adapter, allows you to make a connection, and read tables. So the casing of those table and column names seems to fit there. Convention is used by the planner, which is not doing any name-resolution, just shuffling around references to tables that it has already resolved during the semantic analysis phase. > On Mar 17, 2023, at 1:39 PM, Jonathan Sternberg <[email protected]> wrote: > > Hi, > > I have a system where there will be two conventions. When data moves from > one convention to another, call it from convention A to convention B, the > name of the fields go from all uppercase to lowercase instead. Currently, > in order for the system to operate, we have it so we lie to convention A > and tell convention A that actually the names are lowercase. This has been > causing some issues and the code around convention A would be a lot easier > if it knew the correct casing rather than using the incorrect casing > because of convention B. > > What's the proper way to represent this in Calcite as part of the plan? > > Thank you for any help. > > --Jonathan Sternberg
