[GitHub] [spark] cloud-fan commented on a diff in pull request #39182: [SPARK-41440][CONNECT][PYTHON] Create a new proto message for `RandomSplit`

[GitBox <gi...@apache.org>]

cloud-fan commented on code in PR #39182:
URL: https://github.com/apache/spark/pull/39182#discussion_r1056014283


##########
connector/connect/server/src/main/scala/org/apache/spark/sql/connect/planner/SparkConnectPlanner.scala:
##########
@@ -126,39 +128,37 @@ class SparkConnectPlanner(session: SparkSession) {
     SubqueryAlias(aliasIdentifier, transformRelation(alias.getInput))
   }
 
+  private def transformDeterministicOrder(rel: proto.DeterministicOrder): LogicalPlan = {
+    val input = Dataset.ofRows(session, transformRelation(rel.getInput))
+    // It is possible that the underlying dataframe doesn't guarantee the ordering of rows in its
+    // constituent partitions each time a split is materialized which could result in
+    // overlapping splits. To prevent this, we explicitly sort each input partition to make the
+    // ordering deterministic. Note that MapTypes cannot be sorted and are explicitly pruned out
+    // from the sort order.
+    val sortOrder = input.logicalPlan.output
+      .filter(attr => RowOrdering.isOrderable(attr.dataType))
+      .map(SortOrder(_, Ascending))
+    if (sortOrder.nonEmpty) {
+      Sort(sortOrder, global = false, input.logicalPlan)
+    } else {
+      input.cache()
+      input.logicalPlan
+    }
+  }
+
   /**
    * All fields of [[proto.Sample]] are optional. However, given those are proto primitive types,
    * we cannot differentiate if the field is not or set when the field's value equals to the type
    * default value. In the future if this ever become a problem, one solution could be that to
    * wrap such fields into proto messages.
    */
   private def transformSample(rel: proto.Sample): LogicalPlan = {
-    val input = Dataset.ofRows(session, transformRelation(rel.getInput))
-    val plan = if (rel.getForceStableSort) {

Review Comment:
   we don't need the `force stable sort` anymore?



-- 
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: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org