AngersZhuuuu commented on a change in pull request #30957:
URL: https://github.com/apache/spark/pull/30957#discussion_r551727731
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/BaseScriptTransformationExec.scala
##########
@@ -329,14 +332,45 @@ case class ScriptTransformationIOSchema(
schemaLess: Boolean) extends Serializable {
import ScriptTransformationIOSchema._
- val inputRowFormatMap = inputRowFormat.toMap.withDefault((k) =>
defaultFormat(k))
- val outputRowFormatMap = outputRowFormat.toMap.withDefault((k) =>
defaultFormat(k))
+ val inputRowFormatMap = inputRowFormat.toMap.withDefault(k =>
defaultFormat(k))
+ val outputRowFormatMap = outputRowFormat.toMap.withDefault(k =>
defaultFormat(k))
+
+ val separators = (getByte(inputRowFormatMap("TOK_TABLEROWFORMATFIELD"),
0.toByte) ::
+ getByte(inputRowFormatMap("TOK_TABLEROWFORMATCOLLITEMS"), 1.toByte) ::
+ getByte(inputRowFormatMap("TOK_TABLEROWFORMATMAPKEYS"), 2.toByte) :: Nil)
++
+ (4 to 8).map(_.toByte)
Review comment:
> > why do we need to use different separators for each level?
>
> I have the same question. JSON uses `{}` as the separator and can support
arbitrary nested levels.
IMO, this part should be same with hive for compatibility. If the behavior
is not same, migration will have problem in script. And for JSON, we can add a
Spark's SerDe.
----------------------------------------------------------------
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]
