Support for automatic replacement scans #1244

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

DuckDB has a replacement scan feature in the Python API that allows it to directly access Python variables that are Arrow-based.

This would mean that you wouldn't necessarily need to perform a register_dataframe step in advance; you could just query a variable directly. This would also be temporary; it wouldn't persist the dataframe as a registered object in the session.

Describe the solution you'd like

If a table name is not found among the registered

Is there some way to implement an order of operations for finding tables with a certain name? I.e. inject some code so that if a table x isn't found already-registered, then it can evaluate whether the Python variable x is Arrow-like?

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context

I could implement the Python side of this, but I don't know the DataFusion API well enough to know where to find the hooks or how to connect to them.