maropu commented on a change in pull request #25717: [WIP][SPARK-29013][SQL]
Structurally equivalent subexpression elimination
URL: https://github.com/apache/spark/pull/25717#discussion_r321998484
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/BoundAttribute.scala
##########
@@ -65,6 +65,40 @@ case class BoundReference(ordinal: Int, dataType: DataType,
nullable: Boolean)
}
}
+/**
+ * This bound reference points to a parameterized slot in an input tuple. It
is used in
+ * common sub-expression elimination. When some common sub-expressions have
same structural
+ * but different slots of input tuple, we replace `BoundReference` with this
parameterized
+ * version. The slot position is parameterized and is given at runtime.
+ */
+case class ParameterizedBoundReference(parameter: String, dataType: DataType,
nullable: Boolean)
+ extends LeafExpression {
+
+ override def toString: String = s"input[$parameter,
${dataType.simpleString}, $nullable]"
+
+ override def eval(input: InternalRow): Any = {
+ throw new UnsupportedOperationException(
+ "ParameterizedBoundReference does not implement eval")
+ }
+
+ override def doGenCode(ctx: CodegenContext, ev: ExprCode): ExprCode = {
+ assert(ctx.currentVars == null && ctx.INPUT_ROW != null,
+ "ParameterizedBoundReference can not be used in whole-stage codegen
yet.")
Review comment:
Any barrier to support the whole-stage codegen case?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
