rdblue commented on a change in pull request #3984:
URL: https://github.com/apache/iceberg/pull/3984#discussion_r798138304
##########
File path: core/src/main/java/org/apache/iceberg/io/PositionDeltaWriter.java
##########
@@ -41,6 +41,21 @@
*/
void insert(T row, PartitionSpec spec, StructLike partition);
+ /**
+ * Inserts a new version of an existing row to the provided spec/partition.
+ * <p>
+ * This method allows writers to distinguish new and updated records. The
caller must separately
+ * invoke {@link #delete(CharSequence, long, PartitionSpec, StructLike)} for
the original
+ * row position that is being updated.
+ *
+ * @param row a new version of an existing row
+ * @param spec a new partition spec
+ * @param partition a new partition or null if the spec is unpartitioned
+ */
+ default void update(T row, PartitionSpec spec, StructLike partition) {
Review comment:
If I understand correctly, the problem that this fixes is that the
inserts are expected to be clustered, but we don't have the same expectation
for updates because it's hard to come up with a single sort order that works
for all cases (inserts, updates, and deletes).
So we separate updates that are not expected to be clustered and use a
fanout writer for them?
--
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]
