Highlights
This release brings some meaningful performance improvements to the parsing of complex SQL statements. In files with deeply nested expressions, we have seen up to a 50% reduction on time spent in the parsing phase. These changes are all internal optimisations and have minimal implications for the parser. In a few isolated cases they did highlight inconsistencies in the parsing of literals and so if your use case relies on the specific structure of literal and expression parsing you may find some small differences in how some expressions are parsed.
Additionally this release brings new validation steps to configuration. Layout configuration is now validated on load (and so users with invalid layout configurations may see some of these being caught now) and inline configuration statements in files are also now validated for both their layout rules and for any removed or deprecated settings.
On top of both we've seen dialect improvements to Databricks, PostgreSQL, BigQuery, Snowflake & Athena.
What’s Changed
- Databricks set time zone #5000 @greg-finley
- Terminator inheritance #4981 @alanmcruickshank
- Reduce copying in the parse phase #4988 @alanmcruickshank
- Validate layout configs #4578 #4997 @alanmcruickshank
- Fix handling of keywords for roles in ALTER ROLE statement #4994 @anzelpwj
- BigQuery: fixes parse error on some literals with data type and quoted #4992 @yoichi
- Correct Snowflake
CROSS JOINsyntax #4996 @tunetheweb - Remove broken 'fork me' banner from docs #4989 @greg-finley
- feat: support athena optional WITH ORDINALITY post UNNEST function #4991 @reata