Apache Druid

ChatGpt to the rescue :stuck_out_tongue:
```You need to include it in the list of "aggregations" for the outer query to make sure it is retained in the final result.```
So the following works:
```{
  "queryType": "groupBy",
  "dataSource": {
    "type": "join",
    "left": "table1",
    "right": {
      "type": "query",
      "query": {
        "queryType": "topN",
        "aggregations": [{
            "type": "HLLSketchBuild",
            "name": "tmp_hash",
            "fieldName": "hash_field"
          }]}}
  },
  "filter": {},
  "aggregations": [
    {
      "type": "doubleSum",
      "name": "tmp_hash",
      "fieldName": "b.tmp_hash"
    }
  ],
  ...
  }```

so basically the prefix “b.” is recognised even outside the datasource block.

wow chatgpt should get into our slack channel