[SPARK-26057][SQL] Transform also analyzed plans when dedup references

## What changes were proposed in this pull request?

In SPARK-24865 `AnalysisBarrier` was removed and in order to improve resolution speed, the `analyzed` flag was (re-)introduced in order to process only plans which are not yet analyzed. This should not be the case when performing attribute deduplication as in that case we need to transform also the plans which were already analyzed, otherwise we can miss to rewrite some attributes leading to invalid plans.

## How was this patch tested?

added UT

Please review http://spark.apache.org/contributing.html before opening a pull request.

Closes apache#23035 from mgaido91/SPARK-26057.

Authored-by: Marco Gaido <[email protected]>
Signed-off-by: Wenchen Fan <[email protected]>
(cherry picked from commit b46f75a)
Signed-off-by: Wenchen Fan <[email protected]>

Author: mgaido91

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala

@@ -871,7 +871,7 @@ class Analyzer(
     private def dedupOuterReferencesInSubquery(
         plan: LogicalPlan,
         attrMap: AttributeMap[Attribute]): LogicalPlan = {
-      plan resolveOperatorsDown { case currentFragment =>
+      plan transformDown { case currentFragment =>
         currentFragment transformExpressions {
           case OuterReference(a: Attribute) =>
             OuterReference(dedupAttr(a, attrMap))

sql/core/src/test/scala/org/apache/spark/sql/DataFrameSuite.scala

@@ -2597,4 +2597,29 @@ class DataFrameSuite extends QueryTest with SharedSQLContext {
 
     checkAnswer(swappedDf.filter($"key"($"map") > "a"), Row(2, Map(2 -> "b")))
   }
+
+  test("SPARK-26057: attribute deduplication on already analyzed plans") {
+    withTempView("a", "b", "v") {
+      val df1 = Seq(("1-1", 6)).toDF("id", "n")
+      df1.createOrReplaceTempView("a")
+      val df3 = Seq("1-1").toDF("id")
+      df3.createOrReplaceTempView("b")
+      spark.sql(
+        """
+          |SELECT a.id, n as m
+          |FROM a
+          |WHERE EXISTS(
+          |  SELECT 1
+          |  FROM b
+          |  WHERE b.id = a.id)
+        """.stripMargin).createOrReplaceTempView("v")
+      val res = spark.sql(
+        """
+          |SELECT a.id, n, m
+          |  FROM a
+          |  LEFT OUTER JOIN v ON v.id = a.id
+        """.stripMargin)
+      checkAnswer(res, Row("1-1", 6, 6))
+    }
+  }
 }