fudianchn commented on PR #39613: URL: https://github.com/apache/shardingsphere/pull/39613#issuecomment-5708225064
The schema-drift fallback now restores aggregation-distinct names and labels from the bound projections, matched against returned aliases case-insensitively. It retains newly added backend columns and hides appended derived columns. The previous treatment of an exposed internal alias as an acceptable limitation was incorrect. JDBC `findColumn` also used to bypass the logical label map. It now uses the same map as named value access. Four real H2 query cases check complete names, labels, lookup and values across single/multiple routes, drift/no drift and an explicit alias. Additional tests cover COUNT/SUM/AVG DISTINCT, folded aliases, Proxy query headers and MySQL prepare packet names. The new JDBC regressions fail on the previous public head and pass with the fix; both Proxy consumer regressions also fail before the fix. The implementation-narrative Javadocs on `useExpandedProjections` and `ClientVisibleColumnLayout` have been removed. The cached layout, positional comparison and passthrough behavior remain covered. All 2,041 tests in the three affected modules passed locally, as did Spotless and Checkstyle. The PR description now reflects the current implementation and verification limits. -- 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]
