vladimirg-db opened a new pull request, #49029: URL: https://github.com/apache/spark/pull/49029
### What changes were proposed in this pull request? This is the initial PR for the single-pass Analyzer project. Coverage: - Scans of csv/json/parquet/orc/hive tables and files - `VALUES` - `WHERE` - `LIMIT` - Trivial star expansion - Literals - Aliases - Binary arithmetic - Binary logic Main components: `HybridAnalyzer`: The analysis is now routed between the fixed-point and single-pass by the `HybridAnalyzer`. There are 3 modes: - Fixed-point. In that case `HybridAnalyzer` just calls `executeAndTrack`. This is the default one. - Dual. This one is activated by the `spark.sql.analyzer.singlePassResolver.dualRunWithLegacy` flag and involves a routing logic based on the `ResolverGuard`. First, the query is always analyzed by the fixed-point `Analyzer`, and then, if it's supported by the single-pass, it's analyzed by the `Resolver`. - Single-pass. In that case `HybridAnalyzer` calls the single-pass `Resolver` without any filtering or routing. This is useful for debugging. `Resolver`: Thanks @mihailoale-db and @mihailotim-db for working on that! ### Why are the changes needed? We want to build a new single-pass Analyzer to replace the fixed-point one. SPIP Jira: https://issues.apache.org/jira/browse/SPARK-49834 SPIP: https://docs.google.com/document/d/1dWxvrJV-0joGdLtWbvJ0uNyTocDMJ90rPRNWa4T56Og ### Does this PR introduce _any_ user-facing change? No, the single-pass Analyzer is disabled by default (under the `spark.sql.analyzer.singlePassResolver.enabled` and `spark.sql.analyzer.singlePassResolver.dualRunWithLegacy` flags). ### How was this patch tested? - New test suites - Ran some suites with `spark.sql.analyzer.singlePassResolver.dualRunWithLegacy` enabled ### Was this patch authored or co-authored using generative AI tooling? Copilot. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
