Clarify DateTime32 precision errors

[BharatDeva]

PR checklist:

• Did you run ClangFormat ? Not run locally; the change is small and follows the surrounding style.
• Did you separate headers to a different section in existing community code base ? Not applicable.
• Did you surround proton: starts/ends for new code in existing community code base ? Not applicable; this updates Proton-specific DateTime32 handling that is already in the file.

Please write user-readable short description of the changes:

This makes the datetime32(3) error message more direct.

Before this change, datetime32(3) treated 3 as the first datetime32 argument and reported it as a timezone type error. That is technically accurate from the parser path, but confusing for users because the input looks like a precision argument.

Now, numeric datetime32(...) arguments fail with a clearer message explaining that datetime32 does not support precision and that datetime64 should be used for fractional seconds.

A gtest regression was added for DataTypeFactory::instance().get(datetime32(3)) to check the error code and message.

Validation performed:

git show --check --oneline --summary HEAD

Result:

6b28d1ca Clarify DateTime32 precision errors

I could not run the C++ gtest locally on this Windows checkout because there is no configured build directory or C++ toolchain available in the current shell.

Fixes #816

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[yokofly]

Hi, the changes look good. Please assign the CLA. @BharatDeva #1184 (comment)