Davis-Zhang-Onehouse opened a new issue, #20687: URL: https://github.com/apache/hudi/issues/20687
Part of #20683 (RFC-118, #20684). Defects in the batched BLOB read path (`read_blob()`, `BatchedBlobReadExec`): - `read_blob()` over a join whose two sides have same-named BLOB columns resolves the column by name (`schema.fieldIndex`), so it can return the other side's bytes. - A managed out-of-line BLOB reference holds a path relative to the table; the batched reader is not given the table base path to resolve it. - The storage configuration broadcast to executors is a `HadoopStorageConfiguration`, whose configuration is transient; under Kryo without Hudi's registrator it arrives as null. - A scan that opens one reader per file keeps every exhausted reader reachable from the task's completion listeners until the task ends. Pruned reads of a BLOB column on MOR tables are already handled by master's nested projection; this adds regression tests for them. PR: (linked once opened) -- 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]
