Fixed #33464 -- Resolved output_field for combined numeric expressions with MOD operator.

[David-Wobrock]

ticket-33464

I'm unsure about the second commit, if it makes sense to define as default FloatField + DecimalField => DecimalField.
I'd say that one wants to keep the decimal precision by default, like when doing plain Python arithmetic with a Decimal.

[felixxm]

@David-Wobrock Thanks 👍 I think we should drop the second commit as we will lose precision anyway and users should be aware of that, e.g:

SQL> SELECT MOD(4.6, TO_BINARY_FLOAT(2.5)) FROM DUAL;

MOD(4.6,TO_BINARY_FLOAT(2.5))
-----------------------------
		     2.1E+000

SQL> SELECT CAST(MOD(4.6, TO_BINARY_FLOAT(2.5)) AS NUMBER) FROM DUAL;

CAST(MOD(4.6,TO_BINARY_FLOAT(2.5))ASNUMBER)
-------------------------------------------
				  2.0999999

SQL> SELECT MOD(4.6, 2.5) FROM DUAL;

MOD(4.6,2.5)
------------
	 2.1

[David-Wobrock]

Ah yes @felixxm , that's the caveat I missed. I removed the second commit, thanks 🙏

[felixxm]

@David-Wobrock Thanks 👍