Add sql scalar function to transform value in a row with a single expression

[gertjanal]

Description

Updating an item in an array is possible with the function transform(ARRAY[5, 6], x -> x + 1). This PR adds a new function to update the value for a field in a row. It can be useful when updating multiple rows in an array with a single expression, or when updating an inner value of a nested row:

For a single row:

SELECT transform(
    CAST(ROW('hello', 'world') AS ROW(greeting varchar, planet varchar))",
    'greeting', '', greeting -> concat(greeting, ' or goodbye'));
-- ROW('hello or goodbye', 'world')

For a nested row:

SELECT transform(
    CAST(ROW(ROW('hello'), 'world') as ROW(greeting ROW(text varchar), planet varchar)),
    'greeting',
    CAST(ROW('') as ROW(text varchar)),
    greeting -> transform(
        greeting, 'text', '', text -> concat(text, ' or goodbye')));
-- ROW(ROW('hello or goodbye'), 'world')

SELECT transform(ARRAY[
    cast(ROW('hello', 'world') AS ROW(greeting varchar, planet varchar)),
    cast(ROW('hi', 'mars') AS ROW(greeting varchar, planet varchar))],
    data -> transform(data, 'greeting', '', greeting -> concat(greeting, ' or goodbye')));
-- ARRAY[ROW('hello or goodbye', 'world'), ROW('hi or goodbye', 'mars')]

Additional context and related issues

This PR is a draft as I'd like the opinion of developers:

• Is this a function we want?
• Can the V dummy value be removed and be inferred? I couldn't get it to work without the dummy value; only .argumentType(functionType(UNKNOWN.getTypeSignature(), new TypeSignature("V"))) seemed to match the signature, but I couldn't get UNKNOWN to be updated to the actual value type.
• Is the reading magic with the rawIndex correct? Would it still work for larger data structures?

Release notes

( ) This is not user-visible or is docs only, and no release notes are required.
( x ) Release notes are required. Please propose a release note for me.
( ) Release notes are required, with the following suggested text:

## Functions
* Add sql scalar function to transform value in a row with a single expression

[gertjanal]

This seems to give the right result, but is it correct to get the blocks this way?

[gertjanal]

Added to be consistent with RowType.rowType