Current state

When running matchers for compound (diff-able) data-types, we currently report actual and expected data only.

No diff information is available when running not_to_equal and not_to_contain

Ideal state

When running not_to_equal the report could state:

Actual: refcursor [ count = 10 ] was expected not to equal: refcursor [ count = 4 ]
But all the rows are equal
   [ ... rows from actual go here, as both sides are equal ... ]

When running not_to_contain the report could state:

Actual: refcursor [ count = 10 ] was expected not to contain: refcursor [ count = 4 ]
But it contained:
   [ ... rows from actual that match the expected go here, to show rows that are not meeting the criteria ... ]

In both cases, when using join_by we could probably limit the outcomes to join_by columns as those would be sufficient to indicate the issue.