aokolnychyi commented on code in PR #55518:
URL: https://github.com/apache/spark/pull/55518#discussion_r4429130835
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/RewriteUpdateTable.scala:
##########
@@ -159,14 +164,24 @@ object RewriteUpdateTable extends RewriteRowLevelCommand {
relation: DataSourceV2Relation,
operationTable: RowLevelOperationTable,
assignments: Seq[Assignment],
- cond: Expression): WriteDelta = {
+ cond: Expression,
+ writeAttrs: Seq[AttributeReference],
+ updatedAttrs: Seq[AttributeReference]): WriteDelta = {
val operation = operationTable.operation.asInstanceOf[SupportsDelta]
// resolve all needed attrs (e.g. row ID and any required metadata attrs)
- val rowAttrs = relation.output
val rowIdAttrs = resolveRowIdAttrs(relation, operation)
val metadataAttrs = resolveRequiredMetadataAttrs(relation, operation)
+ val splitUpdates = operation.representUpdateAsDeleteAndInsert
+
+ operation match {
Review Comment:
I'll need to think about this a bit more. Specifically, what's the correct
behavior if we opt in as representing update as delete and insert with narrow
column updates. Keep in mind that Iceberg splits updates but uses synthetic row
IDs... Seems like that would be incompatible with this arrangement then?
--
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]
