numpy unique() returns '<' not supported between float / str instances error - why?

Question

So I have a situation where I have an ID that\'s a concatenation of a three columns, two that are string datatypes and one that\'s an integer.

When I tried to pull,