Union/merge operator and the unique names assumption

[hayesall]

Imagine merging to sets of objects with conflicting names.

For example, two people named "alex":

obj1 = RelationalDataset(pos=['a(alex).'], neg=[], facts=['b(alex).', 'c(alex).'])
obj2 = RelationalDataset(pos=['a(alex).'], neg=[], facts=['b(alex).'])

The interpretation is ambiguous, so should probably result in an error/warning:

>>> obj1 | obj2
UniqueNamesAssumptionError: Found 'alex' in `obj1` and `obj2`

---

If you:

• know these refer to separate entities
• know these refer to the same entities

Then there possible ways to handle this:

>>> with DifferentObjects:
...     obj1 | obj2
RelationalDataset(pos=['a(alex1).', 'a(alex2).'], neg=[], facts=['b(alex1).', 'c(alex1).', 'b(alex2).'])

>>> with SameObjects:
...     obj1 | obj2
RelationalDataset(pos=['a(alex).'], neg=[], facts=['b(alex).', 'c(alex).'])

See also:

• RelationalDataset object union overload #21