Group by and Count in the OSDK is not functioning correctly

Ask the Community
1

I have an ontology object with 158 rows and a single value for source_file_name. When I execute this query:

"groupBy": [
    {
      "type": "exact",
      "field": "source_file_name"
    }
  ],
  "aggregation": [
    {
      "type": "count",
      "name": "count"
    }
  ]

I get this:

{"excludedItems":0,"accuracy":"ACCURATE","data":[{"group":{"source_file_name":"civitas_el_download_20251217_0001.csv"},"metrics":[{"name":"count","value":158.0}]}]}

I should get a count of 1. And when I run this query:

"aggregation": [
    {
      "type": "count",
      "name": "count"
    }
  ]

I get:

{"accuracy":"ACCURATE","data":[{"group":{},"metrics":[{"name":"count","value":158.0}]}]}

Which is the expected result. The only what I can get the count I am looking is with:

"aggregation": [
    {
            "type": "exactDistinct",
            "name": "exactDistinctCount",
            "field": "source_file_name"
        }
  ],
  "where": {
    "type": "or",
    "value": [
      {
        "type": "eq",
        "field": "reviewed",
        "value": false
      },
      {
        "type": "eq",
        "field": "all_checks_pass",
        "value": false
      }
    ]
  }

Which returns:

{"accuracy":"ACCURATE","data":[{"group":{},"metrics":[{"name":"exactDistinctCount","value":1.0}]}]}
2

this aggregation also works, but seems very unintuitive. When you group and count you don’t expect Foundry to do an unwind. An aggregation type of unwind would be more intutitive.

{
  "groupBy": [
    {
      "type": "exact",
      "field": "source_file_name"
    }
  ],
  "aggregation": [
    {
      "type": "exactDistinct",
      "name": "distinctCount",
      "field": "source_file_name"
    }
  ]
}