jackye1995 commented on pull request #1843:
URL: https://github.com/apache/iceberg/pull/1843#issuecomment-738252638


   > how to robustly identify a path in an Identifier (eg not rely solely on 
the presence of `/`)
   
   With all the given constraints, I think we have to rely on some string 
pattern to determine if it is a path or table. Use `/` is probably not a too 
bad choice considering that people won't normally put `/` in the table name. If 
we restrict the path to always be a full file path and do not allow relative 
path, then we can use a URI check (or IO check as you said for extra safety) to 
see if it is a Hadoop table. Is there a use case for relative path for Hadoop 
tables?
   
   And with whatever way we go for the check, I think it should be somewhere in 
the core package, instead of a protected method `isHadoopTable`, so that it can 
be used across compute engines for consistent behavior. One way to go is to add 
it as a method in `HadoopTables`.
   
   > how to correctly handle edge cases like 
catalog.namespace.`file://path/to/table`. What does a namespace mean in this 
case and how do we represent a path as an Identifier or TableIdentifier.
   
   +1 for 
   
   > reject any namespace elements and treat a path as a namespaceless table 
with the full table path as the table/identifier name


----------------------------------------------------------------
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.

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