如何结合 Solr 嵌套文档中的父子节点答案

【问题标题】：how to combine parent and children from Solr nested documents如何结合 Solr 嵌套文档中的父子节点
【发布时间】：2020-01-30 05:19:09
【问题描述】：

我是 lucence、solr 查询的新手，我对如何查询嵌套文档有疑问。

我已将嵌套文档编入索引，如下所示

[
  {
    "id": "1",
    "title": "Solr1",
    "_childDocuments_": [
      {
        "id": "2",
        "title": "Solr2",
        "_childDocuments_": [
          {
            "id": "3",
            "title": "Solr3",
            "_childDocuments_": [
              {
                "id": "4",
                "title": "SolrCloud supports it"
              }
            ],
            "something_else":"irrelevant"
          }
        ],
        "something_else":"irrelevant"
      }
    ],
    "something_else":"irrelevant"
  },
  {
    "id": "5",
    "title": Solr5",
    "_childDocuments_": [
      {
        "id": "6",
        "title": "SolrCloud here as well"
      }
    ]
  }
]

如何搜索标题：SolrCloud，并列出所有孩子的父母？比如

[
  {
    "id": "1",
    "title": "Solr1",
    "_childDocuments_": [
      {
        "id": "2",
        "title": "Solr2",
        "_childDocuments_": [
          {
            "id": "3",
            "title": "Solr3",
            "_childDocuments_": [
              {
                "id": "4",
                "title": "SolrCloud supports it"
              }
            ]
          }
        ]
      }
    ]
  },
  {
    "id": "5",
    "title": Solr5",
    "_childDocuments_": [
      {
        "id": "6",
        "title": "SolrCloud here as well"
      }
    ]
  }
]

列出了文档 4（Sorl1、Solr2、Solr3）和文档 6（Solr5）的所有父项。而且文档的深度不是常数。

【问题讨论】：

标签： solr lucene edismax dismax

【解决方案1】：

我目前的解决方案是按摩数据，将trace添加到原始数据中，这样我就会知道文件来自哪里。比如

[
  {
    "id": "1",
    "title": "Solr1",
    "_childDocuments_": [
      {
        "id": "2",
        "title": "Solr2",
        **"parent_id":"1",**
        **"trace":"Solr1",**
        "_childDocuments_": [
          {
            "id": "3",
            "title": "Solr3",
            **"parent_id":"2",**
            **"trace":"Solr1/Solr2",**
            "_childDocuments_": [
              {
                "id": "4",
                "title": "SolrCloud supports it"
                **"parent_id":"3",**
                **"trace":"Solr1/Solr2/Solr3",**
              }
            ],
            "something_else":"irrelevant"
          }
        ],
        "something_else":"irrelevant"
      }
    ],
    "something_else":"irrelevant"
  },
  {
    "id": "5",
    "title": Solr5",
    "_childDocuments_": [
      {
        "id": "6",
        **"parent_id":"5",**
        **"trace":"Solr5",**
        "title": "SolrCloud here as well"
      }
    ]
  }
]

所以索引后，我可以从结果中知道谁是父文档。

有人可以同意吗？寻找比这更好的解决方案。

【讨论】：