maropu commented on a change in pull request #29900:
URL: https://github.com/apache/spark/pull/29900#discussion_r498056679
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/OptimizeJsonExprs.scala
##########
@@ -38,6 +43,15 @@ object OptimizeJsonExprs extends Rule[LogicalPlan] {
// so `JsonToStructs` might throw error in runtime. Thus we cannot
optimize
// this case similarly.
child
+
+ case g @ GetStructField(jsonToStructs: JsonToStructs, ordinal, _) =>
+ val prunedSchema =
StructType(Seq(jsonToStructs.schema.asInstanceOf[StructType](ordinal)))
Review comment:
nit: we might be able to filter out a schema.length=1 case:
```
case g @ GetStructField(jts @ JsonToStructs(schema: StructType, _, _,
_), ordinal, _)
if schema.length > 1 =>
val prunedSchema = StructType(Seq(schema(ordinal)))
```
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/OptimizeJsonExprs.scala
##########
@@ -38,6 +43,15 @@ object OptimizeJsonExprs extends Rule[LogicalPlan] {
// so `JsonToStructs` might throw error in runtime. Thus we cannot
optimize
// this case similarly.
child
+
+ case g @ GetStructField(jsonToStructs: JsonToStructs, ordinal, _) =>
+ val prunedSchema =
StructType(Seq(jsonToStructs.schema.asInstanceOf[StructType](ordinal)))
+ g.copy(child = jsonToStructs.copy(schema = prunedSchema), ordinal = 0)
+
+ case g @ GetArrayStructFields(jsonToStructs: JsonToStructs, _, _, _, _)
=>
Review comment:
ditto
----------------------------------------------------------------
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:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
