Issues handling Pandas valid column names

[JohnOmernik]

I've noticed that if you have a pandas column that has a - or a space in it (probably other characters too) that Pandas finds acceptable, this can completely break Fugue when doing something as simple as a SELECT * from the Dataframe. Some random thoughts on this:

1. If Pandas supports it as a column name, Fugue should too. We need to find a way to handle these column names.
2. The error messages are exceptions and not are not helpful to a user.
3. In SQL, when someone tries to name a column the same name as a reserved keyword, similar issues happen. In the SQL world, backticks around columns names provides the way to handle many cases here. This wouldn't help SELECT * however, I believe that if we do a SELECT * we should materialize the results with the same columns without issue.
4. The risk of not handling is this user acceptance frustration. If they are reading a CSV that has dashes in the column names, or pulling a large dataset in from parquet files that may have some character that Fugue doesn't support, they will just not use Fugue. Yes, you can rename columns, but from my perspective, if people are using Fugue SQL, they may not have the skills to natively use Pandas to fix the columns in a way that would allow them to run the FSQL. Anything that is an Analysis blocker should be addressed.