bug: create_table/insert wrong context execution for Table

[memeplex]

What happened?

I believe this may be a bug.

import ibis

bq_con = ibis.bigquery.connect(...)
dd_con = ibis.duckdb.connect()

bq_t = bq_con.table("foo")
dd_con.create_table("foo", bq_t.head(10), overwrite=True)

raises BinderException: Binder Error: Catalog "..." does not exist!, where ... is a catalog referenced from the source expression that doesn't exist in the target database and IMO should be resolved against the source database instead.

Also

dd_t = dd_con.create_table("head", schema=bq_t.schema(), overwrite=True)
dd_con.insert("head", bq_t.head(10))

suffers from the same issue.

Related discussion: #11681

What version of ibis are you using?

10.8.2

What backend(s) are you using, if any?

duckdb, bigquery (but can be reproduced with just two duckdb databases)

Relevant log output

Code of Conduct

• I agree to follow this project's Code of Conduct

[memeplex]

I realized something similar was reported in #4800 and then #8115. Even though I would like the feature to be implemented, in the meantime I believe the current behavior is too confusing and should be documented, something like: "note that a ir.Table from a different database is not supported".