"COPY TO" should allow writing to an existing empty folder in S3 bucket

Steps to reproduce - 

1. Create an on-prem cluster with serverless mode.
2. Setup the compute storage seperation.
3. Now load travel-sample bucket on KV
4. create a dataset on travel-sample.
5. Now execute copy to statement, such that the path specified in the statement already exists in the S3 bucket.
6. Following error is thrown

   1. [

      {

      "code": 23073,

      "msg": "Cannot write to a non-empty directory 'test2/' (in line 1, at column 41)",

      "query_from_user": "COPY Default.x TO copytotest AT l1 PATH(\"test2/\");"

      }

      ]

      here test2 folder already exists in S3 bucket, also the folder is empty.