Support optional AS keyword in CTE definitions for Databricks

[funcpp]

Summary

• Add Dialect::supports_cte_without_as() to support dialects where the AS keyword is optional in CTE definitions
• Enable for Databricks and Generic dialects
• Use maybe_parse to disambiguate name (query) from name (columns), following the same pattern as IN subquery/list disambiguation

Databricks allows omitting the AS keyword:

WITH cte (SELECT1) SELECT*FROM cte

Test plan

• Positive tests: CTE without AS, mixed CTEs, column defs without AS
• Negative test: non-supporting dialects reject the syntax
• Existing CTE tests unaffected
• cargo fmt, cargo clippy, full test suite pass

[funcpp]

@iffyio Hi! Could you take a look at these Databricks dialect PRs when you have a chance?

• Support optional AS keyword in CTE definitions for Databricks #2286 (CTE without AS)
• Enable ! as NOT operator for Databricks dialect #2287 (! operator)
• Allow bare columns in GROUPING SETS expressions #2288 (GROUPING SETS bare columns)
• Support multi-column aliases in SELECT items #2289 (multi-column alias)
• Enable numeric-prefix identifiers for Databricks dialect #2290 (numeric-prefix identifiers)

Thanks!

[yoavcloud]

If there's a way to remove some duplications in this function, it would be great. It's not critical, though IMO.

[funcpp]

thanks for the review! I merged the AS and column-defs paths, is it better?

[yoavcloud]

@funcpp, please address the conflicts and see if my comment makes sense. Otherwise, looks good.