Creating secondary indexes on dataset with columnar storage fails with Internal error

Steps to reproduce -

1. Create a 4 node cluster with 2 cbas and 2 KV,N1ql and index nodes.
2. Enable end to end encryption and set it to control.
3. Set default storage format for datasets to column.
4. Restart the analytics cluster.
5. Create a bucket with 3 scopes and 1 collection in each scope.
6. Load data into the collections.
7. Create multiple datasets on different collections.
8. Try creating index on one of the dataset.
9. Following error is seen -

   [

   {

   "code": 25000,

   "msg": "Internal error",

   "query_from_user": "create analytics index JsL on Default.ysr(name:STRING);"

   }

   ]

Error is not seen if index is created on a dataset with row storage format.