Why count(distinct patient_id) is larger than count(patient_id)?

[DongPoJuShi_Dj]

There is an index alias that includes two indexes, and the index structure is as follows:

{
  "scientific_data_group1": {
    "aliases": {
      "scientific_data_group": {}
    },
    "mappings": {
      "properties": {
        "category_id": {
          "type": "keyword"
        },
        "category_name": {
          "type": "keyword"
        },
        "id": {
          "type": "keyword"
        },
        "patient_id": {
          "type": "keyword"
        }
      }
    }
  },
  "scientific_data_group_nested_20230525144327776": {
    "aliases": {
      "scientific_data_group": {}
    },
    "mappings": {
      "properties": {
        "category_id": {
          "type": "keyword"
        },
        "category_name": {
          "type": "keyword"
        },
        "id": {
          "type": "keyword"
        }
        "patient_id": {
          "type": "keyword"
        }
      }
    }
  }
}

Now, when using SQL syntax to query, the result of count(distinct patient_id) is larger than count(patient_id). What could be the possible reason for this?

POST _sql?format=json
{
  "query": """
SELECT count(patient_id),count(distinct patient_id),category_id FROM scientific_data_group
group by category_id
"""
}