That's not possible, because the group_by fields are used to create the document id's. This must be deterministic, because otherwise transform would not be able to update existing documents. The result would e.g. duplicated docs.
We could however relax this restriction for the aggregation section in future. If you use a script in the aggregation part you can use named scripts and change the script later.
What's the reason not to clone the transform instead? If you change essential parts like the group_by, it is conceptually a different transform.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.