Use Case 1 - Using system column ROWID as primary key
I need to Diff two tables (Oracle and Redshift) that don't have true primary key columns. The Oracle table is regularly replicated to Redshift, including the "ROWID" system identifier column, so that column also exists in Redshift though under a different name. When I supply "ROWID" as the key column to the Oracle table, the tool errors out due to not being able to find the column in Oracle's ALL_TAB_COLUMNS.

Use Case 2 - Diffs between Views
It would be useful to also have the ability to diff arbitrary views, or a View vs Table. Right now I don't think this is possible because the tool always tries to look up column types in the information schema.

Describe the solution you'd like
I'm working on a branch that allows you to override the column types so that we can use columns that don't show up in the DB's information schema (eg. ALL_TAB_MODIFICATIONS). Would love your feedback as to whether this is the right solution to the problem.

Describe alternatives you've considered
Another possible solution to Use Case 2 is to allow for passing in the underlying tables that make up a view so that data-diff can look up the column types.