voonhous commented on PR #19029:
URL: https://github.com/apache/hudi/pull/19029#issuecomment-5102633909

   Following up on my own two lists:
   
   - **Durability across engines** -- done. Config doc and the 
`SchemaCompatibilityException` both now say the table schema is corrected but 
the base files aren't, and that they need a clustering/compaction rewrite 
before non-Hudi engines read them correctly.
   - **Flag ignored on some write paths** -- resolved by the redesign. 
`reconcileTimestampLogicalType` sits above the reconcile/no-reconcile split in 
`deduceWriterSchema`, so `MERGE INTO` and `hoodie.avro.schema.validate=true` 
both hit the guard now. 
`testReconcileTimestampLogicalTypeGuardsNonReconcilePath` pins it.
   - **Col stats after a relabel** -- still unexamined.
   
   Also fixed while in there: both the config doc and the exception pointed at 
"the timestamp inspection tool", which doesn't exist in this repo -- the only 
one is in a private repo. They now describe the actual rule (judge the stored 
longs, ~1e12 vs ~1e15, non-overlapping) and point at 
`TimestampLogicalTypeClassifier`.
   
   Still open from the first pass: the cross-zone `assertThrows`, the 
bidirectional-gate question, and the `long -> local-timestamp` doc gap 
wombatu-kun raised.
   


-- 
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]

Reply via email to