Row format: support omitting null byte for non-nullable types

**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**
For primitives, the row format

**Describe the solution you'd like**

When given a datatype that is not nullable, we don't have to write the first byte (`1_u8`) as every value will be valid. This saves some time / memory usage.

We can add a field / constructor to `SortField`.

**Describe alternatives you've considered**

**Additional context**

See slack question/discussion: https://the-asf.slack.com/archives/C01QUFS30TD/p1692631551160779

Start by locating the row-format implementation and the SortField definition or constructors mentioned in the issue. Trace how nullability is encoded for primitive values and identify the existing tests for row serialization. Done means non-nullable types can omit the leading null byte while nullable types retain the current format, with coverage for both cases.

Written by the indexing model from the issue text.