Query aggregation buckets for unique sets of terms in Elasticsearch

Given the following index:

PUT /example
{
  "mappings": {
    "properties": {
      "tags": {
        "type": "keyword"
      }
    }
  }
}

POST example/_bulk
{ "create" : { "_index" : "example" } }
{ "tags" : ["a", "b"] }
{ "create" : { "_index" : "example" } }
{ "tags" : ["c", "d"] }
{ "create" : { "_index" : "example" } }
{ "tags" : ["e"] }
{ "create" : { "_index" : "example" } }
{ "tags" : ["c", "d"] }

I want to aggregate them by the unique set of tags, rather than by all documents that contain the tags. Similar to multi-terms aggregation, but looking at one field. So the response I'm looking for looks like this:

{
  ...
  "aggregations" : {
    "tags" : {
      "doc_count_error_upper_bound" : 0,
      "sum_other_doc_count" : 0,
      "buckets" : [                       
        {
          "key" : ["a", "b"],
          "key_as_string" : "a|b",
          "doc_count" : 1
        },
        {
          "key" : ["c", "d"],
          "key_as_string" : "c|d",
          "doc_count" : 2
        },
        {
          "key" : ["e"],
          "key_as_string" : "e",
          "doc_count" : 1
        }
      ]
    }
  }
}

I know one way to achieve this is to create another field that is a sorted string of all the tags and then do an aggregation on that field, but I just want to know if this is possible. My real use case is a little more complicated, using nested fields, so I'd like to avoid adding a new field.

try this:

PUT test_example
{
  "mappings": {
    "properties": {
      "tags": {
        "type": "keyword"
      }
    }
  }
}

POST _bulk?refresh
{ "create" : { "_index" : "test_example" } }
{ "tags" : ["a", "b"] }
{ "create" : { "_index" : "test_example" } }
{ "tags" : ["c", "d"] }
{ "create" : { "_index" : "test_example" } }
{ "tags" : ["e"] }
{ "create" : { "_index" : "test_example" } }
{ "tags" : ["c", "d"] }

POST test_example/_search
{
  "size": 0,
  "aggregations": {
    "tags": {
      "terms": {
        "script": {
          "source": "String.join('|', params._source.tags)",
          "lang": "painless"
        },
        "collect_mode": "breadth_first",
        "execution_hint": "map"
      }
    }
  }
}

EDIT

If you want to keep "c|d" and "d|c" in the same bucket you can use a custom and you can sort the tags manually using a custom script during the search. Here's an updated query:

POST _bulk?refresh
{ "create" : { "_index" : "test_example2" } }
{ "tags" : ["c", "d"] }
{ "create" : { "_index" : "test_example2" } }
{ "tags" : ["d", "c"] }

POST test_example2/_search
{
  "size": 0,
  "aggregations": {
    "tags": {
      "terms": {
        "script": {
          "source": """
            def sortedTags = params._source.tags.stream().sorted().collect(Collectors.toList());
            String.join('|', sortedTags)
          """,
          "lang": "painless"
        },
        "collect_mode": "breadth_first",
        "execution_hint": "map"
      }
    }
  }
}