Remove test using Tiny/bad-assets-data.csv (and file)

[clizbe]

What and Why

This file was used for testing and is out of date. If we still use it, we should update it. Otherwise, remove.

Possible Drawbacks

No response

Related Issues

No response

[datejada]

We use it to test if the type defined in the data differs from the one specified in the code (e.g., Integer and then having a String in the file). It doesn't need to be updated since the idea is to test that you get an error if you read a file with a bad schema; see the screenshot below:

So, I don't think this is an issue. But, just in case, let's double-check with @suvayu before closing it, who was the person who recommended this test at the beginning 😄

[suvayu]

Indeed, testing for malformed/incorrect input files is a good thing, but I don't think that test actually does that. The test is actually "testing" if the dict holding all the schemas has a schema called "bad_assets_data", which is not really a useful test. In pseudo-code, the test should look like this:

schema = TEM.get_schema("assets_data")
@test_throws <ExceptionType> TIO.create_tbl(con, ...; types = schema)

Ideally you would catch a specific exception, but if that's not possible you should also check the error message. So if the <ExceptionType> is ErrorException, there should be a follow-up like this:

@test_throws r"<regex matching message>" TIO.create_tbl(con, "/path/to/bad_assets_data.csv"; types = schema)

The code duplication is unavoidable due to Julia testing library limitations. You can see this example from TIO.

[suvayu]

Actually, you can drop this test, because this should be tested in TIO, which I think it already is.

EDIT: Maybe the test in TIO can be extended to add a bad data test, I see that now it only tests for success, no failure