tom-s-powell opened a new pull request, #18013: URL: https://github.com/apache/iceberg/pull/18013
Separate normalized file identity from the original URI. Spark now uses `FileIdentifier` for the valid-file side of the orphan-file join, so it shuffles only `path`, `scheme`, and `authority`. ## Why The valid-file side previously used `FileURI`, including `uriAsString`. Matching does not use this duplicate URI string, but Spark still transferred it during the join. This adds network and sort data for tables with many files. ## How it works `FileIdentifier` contains the normalized fields required for matching. `FileURI` extends it and retains the original URI string required for deletion and output. The valid-file dataset now uses `Dataset<FileIdentifier>`, while listed files continue to use `Dataset<FileURI>`. The change is applied to Spark 3.5, 4.0, 4.1, and 4.2. ## Testing - Full remove-orphan action tests for Spark 3.5–4.2 - Core `FileIdentifier` and `FileURI` tests - Core Revapi compatibility check - Flink 2.3 compilation - Spotless checks -- 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]
