Is there a common workaround to have multiple data types in a MAP in Flink SQL?

In general the data structure should remain the same, inside the body is a map with multiple values. No nested fields should be added or anything. I already tried using multiple maps or using a MultiSet and some things like that but unfortunately i did not succeed.
My flink statement looks something like this:

ARRAY[
        ROW(
            ...
            -- Body
            map [
                'Entry1', Entry1,
                'Entry2', Entry2,
                'Entry3', Entry3
            ]
        )
    ] compositeRequest 

These are just dummy entries and not real ones obviously.

My body of the schema looks like this:

{
  "default": null,
  "name": "body",
  "type": [
    "null",
    {
     "type": "map",
     "values": [
       "null",
       "string"
      ]
     }
  ]
}

It can also happen that one of the entries is null, but this is already in my schema.

In general the data structure should remain the same, inside the body is a map with multiple values. No nested fields should be added or anything. I already tried using multiple maps or using a MultiSet and some things like that but unfortunately i did not succeed.