yyanyy commented on a change in pull request #2465:
URL: https://github.com/apache/iceberg/pull/2465#discussion_r612759918
##########
File path: core/src/main/java/org/apache/iceberg/SchemaUpdate.java
##########
@@ -408,11 +450,15 @@ private TableMetadata applyChangesToMapping(TableMetadata
metadata) {
private static Schema applyChanges(Schema schema, List<Integer> deletes,
Map<Integer, Types.NestedField> updates,
Multimap<Integer, Types.NestedField> adds,
- Multimap<Integer, Move> moves) {
+ Multimap<Integer, Move> moves,
+ Set<String> rowIdentifierNames) {
Types.StructType struct = TypeUtil
.visit(schema, new ApplyChanges(deletes, updates, adds, moves))
.asNestedType().asStructType();
- return new Schema(struct.fields());
+ Set<Integer> refreshedRowIdentifiers = rowIdentifierNames.stream()
+ .map(name -> struct.field(name).fieldId())
+ .collect(Collectors.toSet());
+ return new Schema(struct.fields(), refreshedRowIdentifiers);
Review comment:
I think we may want a change to `TableMetadata` to ensure an update to
only row identifier should still create a new schema; I think currently if a
schema update does not change the underlying `struct`, we will consider the
operation a no-op (comparison on `struct` is also why we don't have
`equals`/`hashcode` in `Schema`)
--
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]
