YannByron commented on code in PR #8414:
URL: https://github.com/apache/paimon/pull/8414#discussion_r3511186333
##########
paimon-spark/paimon-spark-common/src/main/scala/org/apache/paimon/spark/catalyst/analysis/PaimonAnalysis.scala:
##########
@@ -102,12 +99,201 @@ class PaimonAnalysis(session: SparkSession) extends
Rule[LogicalPlan] {
}
}
+ private def resolvePaimonWrite(
+ v2WriteCommand: V2WriteCommand,
+ table: DataSourceV2Relation,
+ options: Options,
+ mergeSchemaEnabled: Boolean): LogicalPlan = {
+ val query = stripHiveDynamicPartitionMarker(v2WriteCommand.query)
+ hiveDynamicPartitionColumns(v2WriteCommand.query) match {
+ case Some(dynamicPartitionColumns) if !v2WriteCommand.isByName =>
+ resolveHiveDynamicPartitionWrite(
+ query,
+ table,
+ dynamicPartitionColumns,
+ options,
+ mergeSchemaEnabled)
+ case _ =>
+ v2WriteCommand match {
+ case o: OverwritePartitionsDynamic if !o.isByName =>
+ resolveDynamicPartitionOverwrite(
+ o.copy(query = query),
+ table,
+ options,
+ mergeSchemaEnabled)
+ case _ =>
+ val expected =
+ expectedAttrsForWrite(query, table, options,
v2WriteCommand.isByName)
+ resolveWriteOutput(
+ query,
+ table.name,
+ expected,
+ v2WriteCommand.isByName,
+ mergeSchemaEnabled)
+ }
+ }
+ }
+
+ private def hiveDynamicPartitionColumns(query: LogicalPlan):
Option[Seq[String]] = {
+ query.collectFirst {
+ case PaimonHiveDynamicPartitionQuery(dynamicPartitionColumns, _) =>
+ dynamicPartitionColumns
+ }
+ }
+
+ private def stripHiveDynamicPartitionMarker(query: LogicalPlan): LogicalPlan
= {
+ query.transformDown { case PaimonHiveDynamicPartitionQuery(_, child) =>
child }
+ }
+
+ private def resolveHiveDynamicPartitionWrite(
Review Comment:
`resolveHiveDynamicPartitionWrite` and `resolveDynamicPartitionOverwrite`
are near duplicates. Both methods share the same shape: check
`hiveStyleDynamicPartitionOutput`, and on match do "by-position → by-name",
otherwise a single by-name resolve. Only the way dynamicPartitionColumns is
obtained differs (from the marker vs. from table.partitionKeys()).
--
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]
