任何人都可以请帮助我这个复杂的总和与elasticsearch因为我一直在寻找和尝试了好几天，但似乎没有人有同样的问题，我的.
假设我有这个样本数据：

[
    {
        "id": 4,
        "status": "draft",
        "video": {
            "platforms": [
                {
                    "names": "youtube",
                    "views": 1
                },
                {
                    "names": "dailymotion",
                    "views": 4
                }
            ]
        }
    },
    {
        "id": 2,
        "status": "published",
        "video": {
            "platforms": [
                {
                    "names": "youtube",
                    "views": 2
                },
                {
                    "names": "dailymotion",
                    "views": 10
                }
            ]
        }
    },
    {
        "id": 1,
        "status": "published",
        "video": {
            "platforms": [
                {
                    "names": "youtube",
                    "views": 5
                },
                {
                    "names": "dailymotion",
                    "views": 3
                }
            ]
        }
    }
]

字符串
这是我目前为止所知道的：

{
    "size": 0,
    "aggs": {
        "sum_views": {  
            "nested": {
                "path": "video.platforms"
            },
            "aggs": {
                "platform_terms": {
                    "terms": {
                        "field": "video.platforms.names"
                    },
                    "aggs": {
                        "platform_sum": {
                            "sum": {
                                "field": "video.platforms.views"
                            }
                        }
                    }
                }
            }
        }
    }
}

型
但它们只返回这个：

"aggregations": {
        "sum_views": {
            "doc_count": 3,
            "platform_terms": {
                "doc_count_error_upper_bound": 0,
                "sum_other_doc_count": 0,
                "buckets": [
                    {
                        "key": "youtube",
                        "doc_count": 3,
                        "platform_sum": {
                            "value": 8.0
                        }
                    },
                    {
                        "key": "dailymotion",
                        "doc_count": 3,
                        "platform_sum": {
                            "value": 17.0
                        }
                    }
                ]
            }
        },

型
我正在寻找的是如何从两个平台（单独的），在每个状态的总视图。

[
    {
        "status": "draft",
        "views": [
            {
                "names": "youtube",
                "views": 1
            },
            {
                "names": "dailymotion",
                "views": 4
            }
        ]
    },
    {
        "status": "published",
        "views": [
            {
                "names": "youtube",
                "views": 7
            },
            {
                "names": "dailymotion",
                "views": 13
            }
        ]
    }
]

型
所以他们对youtube和dailymotion进行求和，并按状态对它们进行分组。我尝试过从elasticsearch中进行聚合求和，但我只能分别对youtube和dailymotion中所有视图进行求和，或者对published和draft中的所有视图进行求和。我找不到如何将这两种情况合并结合起来