Gabriel39 opened a new issue, #70316: URL: https://github.com/apache/doris/issues/70316
### Description Continue the native Iceberg encryption reader from #70246 and #70248 by carrying and using each Parquet delete file's independent key and AAD. A plaintext data file can reference encrypted position/equality deletes, and an encrypted data file can reference plaintext deletes. ### Expected scope and verification - Preserve immutable delete encryption metadata through FE scan planning and fresh Thrift descriptors, including the native `position_deletes` metadata table. - Pass data and delete encryption descriptors separately into native Parquet readers. - Keep decrypted position/equality objects and type-converted equality keys out of shared split caches so a warmed file path cannot bypass another split's key validation. - Explicitly reject encrypted ORC and Puffin deletion-vector descriptors and retain the metadata COUNT fence. - Cover all eight plaintext/encrypted data/position/equality combinations, wrong and missing keys, wrong AAD, reused splits, converted equality keys and native metadata-table reads in UT. - Add independently generated Java reference fixtures and SQL regressions for V2 and V2 snapshots upgraded to V3, including real plaintext delete application and continued rejection at encrypted SQL dispatch/COUNT boundaries. Encrypted SQL admission, protected key transport and shared decrypted caches remain separate integration work. This unit should be independently mergeable from authenticated index/Bloom support in #70314. -- 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]
