我的问题是这样的：我有来自不同城市的销售数据：
SHOP NAME | SALES AMOUNT | CITY
shop A, 5000 New York
shop B, 4000 New York
shop C, 3000 New York
shop D, 1800 Boston
shop E, 1500 Boston
shop F, 1300 Boston
现在我想要每个城市的前 2 名销售店铺，然后按销售额的降序排列来自不同城市的前 2 名店铺。所以我们可以得到上面例子的结果：
shop A, 5000 New York
shop B, 4000 New York
shop D, 1800 Boston
shop E, 1500 Boston
我知道他的聚合可以得到前 2 个桶：
POST shop_info/_search
{
"aggs": {
"top_tags": {
"terms": {
"field": "city_id",
"field": "sales",
"size": 3
},
"aggs": {
"top_sales_hits": {
"top_hits": {
"sort": [
{
"sales": {
"order": "desc"
}
}
],
"_source": {
"includes": [ "sales", "city_id" ]
},
"size" : 2
}
}
}
}
}
}
但我不知道如何展平来自不同存储桶的元素：
"buckets": [
{
"key": 1487251891000,
"key_as_string": "2017-02-16 13:31:31",
"doc_count": 10,
"top_sales_hits": {
"hits": {
"total": 10,
"max_score": null,
"hits": [
{
"_index": "shop_info",
"_type": "shop_info",
"_id": "52",
"_score": null,
"_source": {
"ins_tm": "2017-02-16 13:31:31",
"city_id": 12
},
"sort": [
1487251891000
]
},
{
"_index": "shop_info",
"_type": "shop_info",
"_id": "48",
"_score": null,
"_source": {
"ins_tm": "2017-02-16 13:31:31",
"city_id": 12
},
"sort": [
1487251891000
]
}
]
}
}
},
{
"key": 1487251892000,
"key_as_string": "2017-02-16 13:31:32",
"doc_count": 10,
"top_sales_hits": {
"hits": {
"total": 10,
"max_score": null,
"hits": [
{
"_index": "shop_info",
"_type": "shop_info",
"_id": "65",
"_score": null,
"_source": {
"ins_tm": "2017-02-16 13:31:32",
"city_id": 12
},
"sort": [
1487251892000
]
},
{
"_index": "shop_info",
"_type": "shop_info",
"_id": "62",
"_score": null,
"_source": {
"ins_tm": "2017-02-16 13:31:32",
"city_id": 12
},
"sort": [
1487251892000
]
}
]
}
}
},
先感谢您！非常感谢任何提示。