Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/11555#discussion_r55205891 --- Diff: sql/hive/src/main/scala/org/apache/spark/sql/hive/SQLBuilder.scala --- @@ -204,11 +207,70 @@ class SQLBuilder(logicalPlan: LogicalPlan, sqlContext: SQLContext) extends Loggi private def build(segments: String*): String = segments.map(_.trim).filter(_.nonEmpty).mkString(" ") + /** + * Given a seq of qualifiers(names and their corresponding [[AttributeSet]]), transform the given + * expression tree, if an [[Attribute]] belongs to one of the [[AttributeSet]]s, update its + * qualifier with the corresponding name of the [[AttributeSet]]. + */ + private def updateQualifier( + expr: Expression, + qualifiers: Seq[(String, AttributeSet)]): Expression = { + if (qualifiers.isEmpty) { + expr + } else { + expr transform { + case a: Attribute => + val index = qualifiers.indexWhere { + case (_, inputAttributes) => inputAttributes.contains(a) + } + if (index == -1) { + a + } else { + a.withQualifiers(qualifiers(index)._1 :: Nil) + } + } + } + } + + /** + * Finds the outer most [[SubqueryAlias]] nodes in the input logical plan and return their alias + * names and outputSet. + */ + private def findOutermostQualifiers(input: LogicalPlan): Seq[(String, AttributeSet)] = { --- End diff -- I have another alternative. We are facing the same issue everywhere when we add an extra Qualifier or remove an extra Qualifier. How about adding another rule/batch below the existing Batch("Canonicalizer") For example, ```scala Batch("Replace Qualifier", Once, ReplaceQualifier) ``` The rule is simple. We always can get the qualifier from the inputSet if we are doing in bottom up traversal. I did not do a full test last night. Below is the code draft: ```scala object ReplaceQualifier extends Rule[LogicalPlan] { override def apply(tree: LogicalPlan): LogicalPlan = tree transformUp { case plan => plan transformExpressions { case e: AttributeReference => e.withQualifiers(getQualifier(plan.inputSet, e)) } } private def getQualifier(inputSet: AttributeSet, e: AttributeReference): Seq[String] = { inputSet.collectFirst { case a if a.semanticEquals(e) => a.qualifiers }.getOrElse(Seq.empty[String]) } } ```
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org