Good morning, others may speak to their experience in the field when changing data types for individual UDS versus changing the UDS definition, but some things to consider in your example any time users change a data type from Float 32-bit to Integer 32-bit:

• Precision/value loss, i.e. loss of fractional values when going from Float to Integer.
• Range differences, e.g. Unsigned Integer 16-bit allows 0 to 65535, Integer 32-bit allows -2,147,483,648 to 2,147,483,647, etc.
• Instruction/Math calculations may produce different results with different data types.
• HMI/Export mismatches with external device communications that may be expecting the original data type.
• Inconsistency based on individual tag data type changes may be harder to discover later when troubleshooting.

To address your second question, it's allowed for users to handle exception cases. Example: one Motor/Drive reports frequency as Float 32-bit which the UDS was built around, but a replacement unit only responds in 16-bit Integer registers so an exception is made to one UDS instance to handle this by changing data type of the tag.

Thanks for the good question!