Question

我有以下ES内容，基本上是具有针对不同用户的嵌套视频ID的视频列表。

我需要根据特定用户（例如“ B001”）的“ updated_at”日期对列表进行排序。

{ "_index": "videos",
  "_source": {
    "id": 123456,
    "short_title": "First good title",
    "created_at": "2019-05-29T03:19:14",
    "user_list": [
      {
        "user": "A001",
        "video_id": "604214",
        "updated_at": "2019-05-29T03:26:48"
      },
      {
        "user": "B001",
        "video_id": "762001",
        "updated_at": "2019-05-29T06:27:56"
      },
      {
        "user": "C001",
        "video_id": "604218",
        "updated_at": "2019-05-29T06:27:57"
      }
    ]
  }
},
{ "_index": "videos",
  "_source": {
    "id": 127456,
    "short_title": "Second good title",
    "created_at": "2019-05-29T04:19:14",
    "user_list": [
      {
        "user": "B001",
        "video_id": "216001",
        "updated_at": "2019-05-29T06:26:58"
      },
      {
        "user": "A001",
        "video_id": "604218",
        "updated_at": "2019-05-29T06:25:38"
      },
      {
        "user": "C001",
        "video_id": "626001",
        "updated_at": "2019-05-29T06:25:42"
      }
    ]
  }
}

我尝试了以下查询：

GET videos/_search
{
  "query": {
    "bool": {
      "must": [
        {
          "nested": {
            "path": "user_list",
            "filter": {
              "match": {
                "user_list.user": "B001"
              }
            }
          }
        }
      ]
    }
  },
  "from": 0,
  "size": 10,
  "sort": [
    {
      "user_list.updated_at": {
        "order": "asc",
        "nested_path": "user_list",
        "nested_filter": {
          "match": {
            "user_list.user": "B001"
          }
        }
      }
    }
  ]
}

但是我不能得到正确的订单。

预期订单：

{ "_index": "videos",
  "_source": {
    "id": 127456,
    "short_title": "Second good title",
    "created_at": "2019-05-29T04:19:14",
    "user_list": [
      {
        "user": "B001",
        "video_id": "216001",
        "updated_at": "2019-05-29T06:26:58"
      },
      {
        "user": "A001",
        "video_id": "604218",
        "updated_at": "2019-05-29T06:25:38"
      },
      {
        "user": "C001",
        "video_id": "626001",
        "updated_at": "2019-05-29T06:25:42"
      }
    ]
  }
},
{ "_index": "videos",
  "_source": {
    "id": 123456,
    "short_title": "First good title",
    "created_at": "2019-05-29T03:19:14",
    "user_list": [
      {
        "user": "A001",
        "video_id": "604214",
        "updated_at": "2019-05-29T03:26:48"
      },
      {
        "user": "B001",
        "video_id": "762001",
        "updated_at": "2019-05-29T06:27:56"
      },
      {
        "user": "C001",
        "video_id": "604218",
        "updated_at": "2019-05-29T06:27:57"
      }
    ]
  }
}

此处，id为“ 127456”的视频按顺序排在第一位，因为B001用户的“ updated_at”必须按升序排列。

查询中我缺少什么？

Answer 1

映射

data10 = [go.Scatter(
    x = df.portfolio_pnl,
    y = df.sp500_pnl,
    mode = 'markers')]

layout10= dict(
    title = 'Scatter')


updatemenus = list([
    dict(active=-1,
         buttons=list([   
            dict(label = 'S&P 500',
                 method = 'update',
                 args = [{'x': 'df.sp500_pnl'},
                     {}]),
            dict(label = 'Portfolio',
                 method = 'update',
                 args = [{'x': df.portfolio_pnl},
                         {}]),

        ]),
    ),
])

layout10['updatemenus'] = updatemenus

plot({
    'data' : data10, 
    'layout' : layout10,
})

查询

PUT videos/_mapping
{

  "properties" : {
      "user_list" : {
        "type" : "nested",
        "properties" : {
          "user" : {
            "type" : "text"
          },
          "video_id" : {
            "type" : "text"
          },
          "updated_at" : {
            "type" : "date"
          }
        }
      },
      "id" : {
        "type" : "integer"
      },
      "short_title" : {
        "type" : "text"
      },
      "created_at" : {
        "type" : "date"
      }
    }
}

结果

GET videos/_search
{
  "query": {
    "bool": {
      "must": [
        {
          "nested": {
            "path": "user_list",
            "query": {
              "bool": {
                "filter": {
                  "match": {
                    "user_list.user": "B001"
                  }
                }
              }
            }
          }
        }
      ]
    }
  },
  "from": 0,
  "size": 10,
  "sort": [
    {
      "user_list.updated_at": {
        "order": "desc",
        "nested_path": "user_list",
        "nested_filter": {
          "match": {
            "user_list.user": "B001"
          }
        }
      }
    }
  ]
}

我正在使用弹性搜索7.0。查询与您的查询相同，我按降序排序 123456作为“ user”：“ B001”，“ updated_at”：“ 2019-05-29T06：27：56”首先出现大于“用户”：“ B001”，“ updated_at”：“ 2019-05-29T06：26：58”

Elasticsearch如何使用嵌套条件进行排序

1 个答案: