Fix: Resolve error TypeError "unhashable type list" (2nd pass)

[aaronsteers]

A fix for #244.

This is smaller scope than:

• Fix!: Handle cases of anyOf(string, object) by casting to object/JSON/VARIANT #246

In that PR, we modify the failover type to be object/JSON/VARIANT for cases of anyOf(string, object). In this case, we don't change anything except to make sure that the expected exception is raised - and this triggers the expected failsafe behavior, without changing it.

[bindipankhudi]

lgtm! 👍

[aaronsteers]

Noting here for follow-up, I'm not getting consistent rendering across database platforms when I try to validate records converted from object type to string type. The differences are minor, but enough to break tests significantly.

E.g. one database will write the string '{"nested_column": "nested_value"}' and another will write '{"nested_column":"nested_value"}'. I'm simplifying test inputs so we can prioritize the known fix to type resolution itself, and we can come back to this in:

• 💡 Feature Request: Provide more robust handling of anyOf(string, object) scenarios #253