kbendick commented on a change in pull request #1429:
URL: https://github.com/apache/iceberg/pull/1429#discussion_r484601139
##########
File path:
core/src/main/java/org/apache/iceberg/BaseMetastoreTableOperations.java
##########
@@ -75,7 +76,14 @@ public TableMetadata refresh() {
try {
doRefresh();
} catch (NoSuchTableException e) {
- LOG.warn("Could not find the table during refresh, setting current
metadata to null", e);
+ // Adjust log level according to the type of exception, as it might be
intentional to call
+ // the method without determining the type of table in prior (like
SparkSessionCatalog),
+ // and in such case it may not be an error case if the table is not an
Iceberg table.
+ if (e instanceof NoSuchIcebergTableException) {
Review comment:
Oh. I wasn't also considering all of the various `Catalog` classes,
which don't implement `TableOperations` but rely on them pretty heavily. My
question of when is it useful to have an abstract class that is not extended
still stands (question 1 above), but I can see now more of the places where
this code likely comes into play. Thanks so much for helping me get up to speed
on the parts of codebase that I'm less familiar with. As I gain more practical
experience using Iceberg, I'm sure it will become much more clear to me.
Fortunately, I've finally gotten it all running on a modified k8s cluster
somewhere for my own personal practice. 👍
----------------------------------------------------------------
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]
