Github user iyerr3 commented on the issue:
https://github.com/apache/madlib/pull/325
> I agree, this is not a great solution. Casting the operators makes it
especially awkward to use. However, we have to consider the following case. If
a module has multiple test files like `graph` and if they re-use the same table
names like `vertex`, then we have to drop them before re-creating.
You mean if there are multiple tests in a single file, then the 2nd test
needs to drop the table created by the 1st test?
That `drop` statement should be ok since it will drop from the
`install-check` schema and not from the `madlib` schema since we've already
created a table in the `install-check` schema and it is first in the search
path.
We could also schema qualify the first drop statement with the
install-check schema name. This name can be put into a variable by the
preprocessor and then used in the install-check. This would be less
inconvenient since it will only be for the drop statements (and potentially
only for the first drop statement).
---