szehon-ho opened a new issue, #5163:
URL: https://github.com/apache/iceberg/issues/5163

   **Background**: #3851 and further prs added support to register a new table 
in catalog pointing to existing metadata.json.  It may be good to support 
setting an existing table in catalog to another metadata.json.
   
   **Justification**:  There have been instances in the past where such a 
method would have been welcome, the workaround unfortunately is to manually 
update the catalog (ie, update Hive Metastore directly).
   
   1. Disaster recovery where the catalog backup has a table pointing to an 
older metadata.json
   2. Bugs where table becomes in a bad state due to metadata modification (ie, 
partition field dropping bugs that corrupted the table: see 
https://apache-iceberg.slack.com/archives/C025PH0G1D4/p1654877950329549, ref: 
https://github.com/trinodb/trino/issues/8284
   3. Catalog consistency problems where metadata.jsons get overriden from each 
other (ie #5036, or some custom catalogs )
   
   **Ideas:**
   
   1. Add a force option to catalog::registerTable 
   2. Add a different method to catalog (reset table?)
   
   **Concerns:**
   
   This could of course be dangerous, but metadata.json does seem a 
user-exposed concept already in catalog API and in some utils.
   
   We could potentially run some kind of table consistency check (explore all 
the reachable graph and even historic metadata graph to validate that the new 
table metadata is consistent.
   


-- 
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]

Reply via email to