EnricoMi commented on code in PR #37407:
URL: https://github.com/apache/spark/pull/37407#discussion_r972844049


##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala:
##########
@@ -869,26 +873,55 @@ class Analyzer(override val catalogManager: 
CatalogManager)
     def apply(plan: LogicalPlan): LogicalPlan = 
plan.resolveOperatorsWithPruning(
       _.containsPattern(UNPIVOT), ruleId) {
 
-      // once children and ids are resolved, we can determine values, if non 
were given
-      case up: Unpivot if up.childrenResolved && up.ids.forall(_.resolved) && 
up.values.isEmpty =>
-        up.copy(values = up.child.output.diff(up.ids))
+      // once children are resolved, we can determine values from ids and vice 
versa
+      // if only either is given
+      case up: Unpivot if up.childrenResolved &&
+        up.ids.exists(_.forall(_.resolved)) && up.values.isEmpty =>
+        up.copy(values =
+          Some(
+            up.child.output.diff(up.ids.get.flatMap(_.references))

Review Comment:
   Yes, this implements `.unpivot(Array, String, String)`, when no value are 
given, similar to when no ids are given in SQL.
   
   This is the implementation of
   ```
      * Note: A column that is referenced by an id column expression is 
considered an id column itself.
      * For instance `$"id" * 2` references column `id`, so `id` is considered 
an id column and not a
      * value column
   ```
   
   I thought, the following would be annoying:
   ```
   scala> val df = spark.range(5).select($"id", ($"id"*10).as("val"))
   scala> df.show()
   +---+---+
   | id|val|
   +---+---+
   |  0|  0|
   |  1| 10|
   |  2| 20|
   |  3| 30|
   |  4| 40|
   +---+---+
   
   df.unpivot(Array($"id" * 2), "col", "val").show()
   +--------+---+---+
   |(id * 2)|col|val|
   +--------+---+---+
   |       0|id|  0|
   |       0|val|  0|
   |       2|id| 1|
   |       2|val| 10|
   |       4|id 2|
   |       4|val| 20|
   |       6|id 3|
   |       6|val| 30|
   |       8|id 4|
   |       8|val| 40|
   +--------+---+---+
   ```
   
   Of course, that id manipulation can be done before `unpivot` to 
"materialize" it as a reference:
   ```
   df.withColumn("id", $"id" * 2).unpivot(Array($"id"), "col", "val").show()
   +---+---+---+
   | id|col|val|
   +---+---+---+
   |  0|val|  0|
   |  2|val| 10|
   |  4|val| 20|
   |  6|val| 30|
   |  8|val| 40|
   +---+---+---+
   ```
   
   Happy to remove that complexity.



-- 
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]

Reply via email to