danielcweeks commented on code in PR #12250:
URL: https://github.com/apache/iceberg/pull/12250#discussion_r1974017826
##########
spark/v3.5/spark/src/main/java/org/apache/iceberg/spark/source/SparkPositionDeletesRewrite.java:
##########
@@ -200,7 +207,8 @@ static class PositionDeletesWriterFactory implements
DataWriterFactory {
StructType dsSchema,
int specId,
StructLike partition,
- Map<String, String> writeProperties) {
+ Map<String, String> writeProperties,
+ int formatVersion) {
Review Comment:
I would think that if you capture the format version when you create the
SerializableMetadataTable [like we do with name
here](https://github.com/apache/iceberg/blob/22d14d5407356989415ccabd50071392c3c1076c/core/src/main/java/org/apache/iceberg/SerializableTable.java#L445)
and it would be available for all metadata tables (not just positional deletes)
I believe @aokolnychyi had a [similar
comment](https://github.com/apache/iceberg/pull/11620#discussion_r1859414335)
and I'm not sure I follow @RussellSpitzer concern (other than we already have
multiple places to check to find what format version we're working with).
In general, it would be nice if a table acts like a table and we can get the
format version we're working against. This just feels weird because we're
updating the methods to pass through info we should have from the table.
--
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]
