Enable unqualified table reads

[C0urante]

The source connector uses fully-qualified table names when issuing SELECT queries, and requires that the set of tables that it reads from does not contain any two tables with the same unqualified name.

This makes sense for most use cases, but is unnecessarily restrictive when, e.g., Postgres search paths are used in the source database.

These can be especially useful if there is a single unqualified table name (or set of table names) that should be read from the source database, but the schema they reside in are constantly changing. A search path can be used to automatically direct queries (that use unqualified table names) to the correct schema without requiring any changes to the connector configuration.

In order to support this type of use case, we can allow the connector to be configured to use unqualified table names when reading from the source database. This would also require us to skip the check we currently perform for duplicate unqualified table names (since that's exactly what we'd be facilitating with this feature).

[jeqo]

Fixed by #223