Clean up null checking

[dsaxton]

From the discussion in #31939 it was pointed out that there are places where we check for nullness in an unnecessarily verbose way, e.g.: https://github.com/pandas-dev/pandas/blob/master/pandas/_libs/hashtable_class_helper.pxi.in#L1036.

Ideally we should just be able to use checknull from pandas/_libs/missing.pyx, but would need to make sure this catches all cases, such as decimal.Decimal("nan") (which I think isn't included at the moment).

[jbrockmendel]

closed by #45071