[BUG] Incorrect behavior when parsing CSV timestamps with out-of-range year with Spark 3.2.0

[andygrove]

Describe the bug
PR #4938 improves timestamp support when reading CSV sources but does not match Spark's behavior since 3.2.0 where there is validation on the date components. For example, Spark throws errors such as ValueError: year 32766 is out of range.

Steps/Code to reproduce bug
Update csv_test.py to remove the XFAIL from simple_int_values that references this issue.

Expected behavior
We should add validation logic to match Spark's behavior.

Additional context
None

[andygrove]

This error was actually due to Python's limitation of max year 9999