Summary

Add support for parenthesized multi-column aliases in SELECT items, as defined in the Spark SQL grammar:

namedExpression
    : expression (AS? (name=errorCapturingIdentifier | identifierList))?
    ;

identifierList
    : LEFT_PAREN identifierSeq RIGHT_PAREN
    ;

This enables syntax like:

SELECT stack(2, 'a', 'b', 'c', 'd') AS (col1, col2)

Changes

• Add SelectItem::ExprWithAliases variant for multi-column aliases
• Add Dialect::supports_select_item_multi_column_alias(), enabled for Databricks and Generic dialects
• Parse AS (ident, ident, ...) when the dialect supports it

Context

While not documented in the Databricks SQL reference, this syntax is part of the Spark SQL grammar that Databricks implements. Verified to execute successfully on Databricks Runtime:

SELECT stack(2, 'a', 'b', 'c', 'd') AS (col1, col2)
-- Returns:
-- col1  col2
-- a     b
-- c     d

Test plan

• Round-trip test: SELECT stack(...) AS (col1, col2) with and without FROM
• Negative test: non-supporting dialects reject the syntax
• Existing tests unaffected
• cargo fmt, cargo clippy, full test suite pass