tomvanbussel commented on code in PR #38979:
URL: https://github.com/apache/spark/pull/38979#discussion_r1044332219
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/planner/SparkConnectPlanner.scala:
##########
@@ -380,7 +396,26 @@ class SparkConnectPlanner(session: SparkSession) {
}
val attributes = structType.toAttributes
val proj = UnsafeProjection.create(attributes, attributes)
- new logical.LocalRelation(attributes, rows.map(r => proj(r).copy()).toSeq)
+ val relation = logical.LocalRelation(attributes, rows.map(r =>
proj(r).copy()).toSeq)
+
+ if (rel.hasDatatype || rel.hasDatatypeStr) {
+ // rename columns and update datatypes
+ val schema = if (rel.hasDatatype) {
+ DataTypeProtoConverter
+ .toCatalystType(rel.getDatatype)
+ .asInstanceOf[StructType]
+ } else {
+ parseDatatypeString(rel.getDatatypeStr)
+ .asInstanceOf[StructType]
+ }
+ Dataset
+ .ofRows(session, logicalPlan = relation)
+ .toDF(schema.names: _*)
Review Comment:
Will it work, if we use `val proj = UnsafeProjection.create(attributes,
inferredSchema .toAttributes)` instead? We will likely also have to add some
validation, similar to validation that pyspark currently performs using
`_make_type_verifier`.
##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/planner/SparkConnectPlanner.scala:
##########
@@ -380,7 +396,26 @@ class SparkConnectPlanner(session: SparkSession) {
}
val attributes = structType.toAttributes
val proj = UnsafeProjection.create(attributes, attributes)
- new logical.LocalRelation(attributes, rows.map(r => proj(r).copy()).toSeq)
+ val relation = logical.LocalRelation(attributes, rows.map(r =>
proj(r).copy()).toSeq)
+
+ if (rel.hasDatatype || rel.hasDatatypeStr) {
+ // rename columns and update datatypes
+ val schema = if (rel.hasDatatype) {
+ DataTypeProtoConverter
+ .toCatalystType(rel.getDatatype)
+ .asInstanceOf[StructType]
+ } else {
+ parseDatatypeString(rel.getDatatypeStr)
+ .asInstanceOf[StructType]
+ }
+ Dataset
+ .ofRows(session, logicalPlan = relation)
+ .toDF(schema.names: _*)
Review Comment:
Will it work, if we use `val proj = UnsafeProjection.create(attributes,
inferredSchema .toAttributes)` instead? We will likely also have to add some
validation that is similar to validation that pyspark currently performs using
`_make_type_verifier`.
--
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]
