sql - SQL JOIN 查询中的重复条目-6ren

sql - SQL JOIN 查询中的重复条目

转载作者：行者123 更新时间：2023-11-29 13:23:59

我有三个表posts、comments 和users。 posts 表包括两种类型的帖子，question 和answer。对问题和答案进行评论。我的目的是获得一个带有评论、答案和答案评论的问题。我还需要用户表中的用户名作为我获取的每个问题、答案和评论的作者。我正在使用 Postgres 9.5，并使用 json_agg() 函数。

虽然我需要的示例输出应该类似于以下第一个，但我得到了重复的条目。

我在这里缺少什么？正确的 group by 子句可能是。或者用他们的评论收集答案的子查询不是这样做的方法。当我从评论表上的帖子中注释掉左连接时，我得到了想要的结果，但没有对问题发表评论。此外，当我取消包含子查询的左连接时，我得到了预期的非重复结果，而且这也不是我想要的完整数据集。这些是我迄今为止为解决我的问题而收集的东西。

我需要什么:

[
  {
    "post_id": "10",
    "created_at": "2016-05-10T00:16:54.469Z",
    "post_type": "question",
    "post_title": "qwerty",
    "post_text": "asdasd asda sdasd",
    "post_author_id": 1,
    "author": "isikfsc",
    "parent_post_id": null,
    "is_accepted": null,
    "acceptor_id": null,
    "answers": [
      {
        "post_id": 17,
        "created_at": "2016-05-10T04:58:56.350229",
        "post_type": "answer",
        "post_title": null,
        "post_text": "222asda dasdad asdada",
        "post_author_id": 1,
        "author": "isikfsc",
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 5,
            "created_at": "2016-05-10T10:56:30.220128",
            "text": "qweqwe",
            "author_id": 1,
            "author": "isikfsc",
            "parent_post_id": 17
          },
          {
            "id": 8,
            "created_at": "2016-05-10T11:00:00.182991",
            "text": "sasasd",
            "author_id": 1,
            "author": "isikfsc",
            "parent_post_id": 17
          }
        ]
      },
      {
        "post_id": 14,
        "created_at": "2016-05-10T04:19:19.005556",
        "post_type": "answer",
        "post_title": null,
        "post_text": "asdasdasdasd",
        "post_author_id": 1,
        "author": "isikfsc",
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 2,
            "created_at": "2016-05-10T05:25:34.671008",
            "text": "qeqweqwe",
            "author_id": 1,
            "author": "isikfsc",
            "parent_post_id": 14
          }
        ]
      }
    ],
    "comments": [
        {
          "id": 1,
          "created_at": "2016-05-10T10:56:30.220128",
          "text": "qweqwe",
          "author_id": 1,
          "author": "isikfsc",
          "parent_post_id": 10
        },
        {
          "id": 4,
          "created_at": "2016-05-10T11:00:00.182991",
          "text": "sasasd",
          "author_id": 1,
          "author": "isikfsc",
          "parent_post_id": 10
        }
    ]
  }
]

我的查询是:

SELECT
    q.*,
    json_agg(ac.*) AS answers,
    json_agg(c.*) AS comments --comments on posts of post_id questions
FROM posts q

LEFT JOIN 
    (
        SELECT
            a.*,
            json_agg(c.*) AS comments -- comments on posts of post_id answers
        FROM posts a
        LEFT JOIN comments c
        ON a.post_id = c.parent_post_id

        GROUP BY a.post_id
    ) ac
ON q.post_id = ac.parent_post_id

LEFT JOIN comments c
ON q.post_id = c.parent_post_id

WHERE q.post_id = 10
GROUP BY q.post_id

我得到的:

[
  {
    "post_id": "10",
    "created_at": "2016-05-10T00:16:54.469Z",
    "post_type": "question",
    "post_title": "qwerty",
    "post_text": "asdasd asda sdasd",
    "post_author_id": 1,
    "parent_post_id": null,
    "is_accepted": null,
    "acceptor_id": null,
    "answers": [
      {
        "post_id": 17,
        "created_at": "2016-05-10T04:58:56.350229",
        "post_type": "answer",
        "post_title": null,
        "post_text": "222asda dasdad asdada",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 5,
            "created_at": "2016-05-10T10:56:30.220128",
            "text": "qweqwe",
            "author_id": 1,
            "parent_post_id": 17
          },
          {
            "id": 8,
            "created_at": "2016-05-10T11:00:00.182991",
            "text": "sasasd",
            "author_id": 1,
            "parent_post_id": 17
          }
        ]
      },
      {
        "post_id": 17,
        "created_at": "2016-05-10T04:58:56.350229",
        "post_type": "answer",
        "post_title": null,
        "post_text": "222asda dasdad asdada",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 5,
            "created_at": "2016-05-10T10:56:30.220128",
            "text": "qweqwe",
            "author_id": 1,
            "parent_post_id": 17
          },
          {
            "id": 8,
            "created_at": "2016-05-10T11:00:00.182991",
            "text": "sasasd",
            "author_id": 1,
            "parent_post_id": 17
          }
        ]
      },
      {
        "post_id": 17,
        "created_at": "2016-05-10T04:58:56.350229",
        "post_type": "answer",
        "post_title": null,
        "post_text": "222asda dasdad asdada",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 5,
            "created_at": "2016-05-10T10:56:30.220128",
            "text": "qweqwe",
            "author_id": 1,
            "parent_post_id": 17
          },
          {
            "id": 8,
            "created_at": "2016-05-10T11:00:00.182991",
            "text": "sasasd",
            "author_id": 1,
            "parent_post_id": 17
          }
        ]
      },
      {
        "post_id": 17,
        "created_at": "2016-05-10T04:58:56.350229",
        "post_type": "answer",
        "post_title": null,
        "post_text": "222asda dasdad asdada",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 5,
            "created_at": "2016-05-10T10:56:30.220128",
            "text": "qweqwe",
            "author_id": 1,
            "parent_post_id": 17
          },
          {
            "id": 8,
            "created_at": "2016-05-10T11:00:00.182991",
            "text": "sasasd",
            "author_id": 1,
            "parent_post_id": 17
          }
        ]
      },
      {
        "post_id": 14,
        "created_at": "2016-05-10T04:19:19.005556",
        "post_type": "answer",
        "post_title": null,
        "post_text": "asdasdasdasd",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 2,
            "created_at": "2016-05-10T05:25:34.671008",
            "text": "qeqweqwe",
            "author_id": 1,
            "parent_post_id": 14
          }
        ]
      },
      {
        "post_id": 14,
        "created_at": "2016-05-10T04:19:19.005556",
        "post_type": "answer",
        "post_title": null,
        "post_text": "asdasdasdasd",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 2,
            "created_at": "2016-05-10T05:25:34.671008",
            "text": "qeqweqwe",
            "author_id": 1,
            "parent_post_id": 14
          }
        ]
      },
      {
        "post_id": 14,
        "created_at": "2016-05-10T04:19:19.005556",
        "post_type": "answer",
        "post_title": null,
        "post_text": "asdasdasdasd",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 2,
            "created_at": "2016-05-10T05:25:34.671008",
            "text": "qeqweqwe",
            "author_id": 1,
            "parent_post_id": 14
          }
        ]
      },
      {
        "post_id": 14,
        "created_at": "2016-05-10T04:19:19.005556",
        "post_type": "answer",
        "post_title": null,
        "post_text": "asdasdasdasd",
        "post_author_id": 1,
        "parent_post_id": 10,
        "is_accepted": null,
        "acceptor_id": null,
        "comments": [
          {
            "id": 2,
            "created_at": "2016-05-10T05:25:34.671008",
            "text": "qeqweqwe",
            "author_id": 1,
            "parent_post_id": 14
          }
        ]
      }
    ],
    "comments": [
      {
        "id": 1,
        "created_at": "2016-05-10T05:25:28.200327",
        "text": "asadasdad",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 4,
        "created_at": "2016-05-10T10:25:23.381177",
        "text": "werwer",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 1,
        "created_at": "2016-05-10T05:25:28.200327",
        "text": "asadasdad",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 4,
        "created_at": "2016-05-10T10:25:23.381177",
        "text": "werwer",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 1,
        "created_at": "2016-05-10T05:25:28.200327",
        "text": "asadasdad",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 4,
        "created_at": "2016-05-10T10:25:23.381177",
        "text": "werwer",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 1,
        "created_at": "2016-05-10T05:25:28.200327",
        "text": "asadasdad",
        "author_id": 1,
        "parent_post_id": 10
      },
      {
        "id": 4,
        "created_at": "2016-05-10T10:25:23.381177",
        "text": "werwer",
        "author_id": 1,
        "parent_post_id": 10
      }
    ]
  }
]

最佳答案

一旦所有参与方都加入，分组就会发生，因此聚合将取决于生成的基数。加入带有答案和评论的帖子会导致它们之间的完全连接，从而复制所有值。它们需要分开并单独执行，您可以采用以下一种方法:

SELECT
    q.*,
    (SELECT json_agg(ac.*)
     FROM (
       SELECT a.*, json_agg(c.*) AS comments
       FROM posts a
       LEFT JOIN comments c ON (a.post_id = c.parent_post_id)
       WHERE a.parent_post_id = q.post_id
       GROUP BY a.post_id
       ) ac
    ) AS answers,
    json_agg(c.*) AS comments --comments on posts of post_id questions
FROM posts q
LEFT JOIN comments c ON (q.post_id = c.parent_post_id)
WHERE q.post_id = 10
GROUP BY q.post_id;

或者:

SELECT q.*, qa.answers, qc.comments
FROM posts q
LEFT JOIN (
  SELECT ac.parent_post_id, json_agg(ac.*) AS answers
  FROM (
    SELECT ac.*, json_agg(c.*) AS comments
    FROM posts ac
    LEFT JOIN comments c ON (c.parent_post_id = ac.post_id)
    GROUP BY ac.post_id
    ) ac
  GROUP BY ac.parent_post_id
  ) qa ON (qa.parent_post_id = q.post_id)
LEFT JOIN (
  SELECT c.parent_post_id, json_agg(c.*) AS comments
  FROM comments c
  GROUP BY c.parent_post_id
  ) qc ON (qc.parent_post_id = q.post_id)
WHERE q.post_id = 10;

关于sql - SQL JOIN 查询中的重复条目，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/37151721/

文章推荐： node.js - Postgres 和 Sequelize - 无法读取未定义的属性 'name'

文章推荐： postgresql - Dancer2 Auth::Extensible 不接受散列密码

文章推荐： PHP htmlentities 不工作？输出是？

文章推荐： json - 是否可以针对此事创建查询？

join - 从一个“join”表到另一个“join”表的SqlAlchemy关系
我正在测试设置SQLAlchemy以映射现有数据库。这个数据库是很久以前自动建立的，它是由我们不再使用的先前的第三方应用程序创建的，因此 undefined 某些预期的事情，例如外键约束。该软件将管理
mysql - INNER JOIN、LEFT JOIN、RIGHT JOIN 和 FULL JOIN 有什么区别？
这个问题在这里已经有了答案: What is the difference between "INNER JOIN" and "OUTER JOIN"? (28 个答案) 关闭 7 年前。 INNE
mysql - INNER JOIN、LEFT JOIN、RIGHT JOIN 和 FULL JOIN 有什么区别？
这个问题在这里已经有了答案: What is the difference between "INNER JOIN" and "OUTER JOIN"? (29 个回答) 关闭7年前. INNER J
join - Hive:LEFT JOIN 与 JOIN 在 ON 子句中使用过滤器给出不同的结果
假设有两个表: table1.c1 table1.c2 1 1 A 2 1 B 3 1 C 4 2
join - Hive:LEFT JOIN 与 JOIN 在 ON 子句中使用过滤器给出不同的结果
假设有两个表: table1.c1 table1.c2 1 1 A 2 1 B 3 1 C 4 2
数据库Left join , Right Join, Inner Join 的相关内容，非常实用
一.先看一些最简单的例子例子 Table A aid adate 1 a1 2&nb
SQL 外链接操作小结 inner join left join right join
数据库操作语句 7. 外连接——交叉查询 7.1 查询 7.2 等值连接 7.3 右外
ruby-on-rails - :joins | change behavior inner join to left join
我有两个表 'users' 和 'lms_users' class LmsUser belongs_to :user end class User has_one :lms_user
ruby-on-rails - 首先使用 `joins()` 进行 INNER JOIN 然后是下一个表的 LEFT JOIN
我试图避免在 Rails 中对我的 joins 进行字符串插值，因为我注意到将查询器链接在一起时灵活性会降低。也就是说，我觉得 joins(:table1) 比 joins('inner join
ruby-on-rails - Rails ActiveRecord :joins with LEFT JOIN instead of INNER JOIN
我有这个代码 User.find(:all, :limit => 10, :joins => :user_points, :select => "users.*, co
join - Doctrine join 绕过延迟加载
我刚刚开始探索 Symfony2，我很惊讶它拥有如此多的强大功能。我开始做博客教程在: http://tutorial.symblog.co.uk/ 但使用的是 2.1 版而不是 2.0 我的问题是我
SQL JOIN 和不同类型的 JOIN
什么是 SQL JOIN什么是不同的类型？最佳答案插图来自 W3schools : 关于SQL JOIN 和不同类型的 JOIN，我们在Stack Overflow上找到一个类似的问题： http
join - Hive Join 返回零记录
我有两个 Hive 表，我正在尝试加入它们。这些表没有被任何字段聚集或分区。尽管表包含公共(public)键字段的记录，但连接查询始终返回 0 条记录。所有数据类型都是“字符串”数据类型。连接查询很
join - solr join - 返回父子文档
我正在使用 Solr 的(4.0.0-beta)连接功能来查询包含具有父/子关系的文档的索引。连接查询效果很好，但我只能在搜索结果中获得父文档。我相信这是预期的行为。但是，是否有可能在搜索结果中同时
join - 三向关联查询/has_many :through/join
我正在使用可用的指南/api/书籍自学 Rails，但我无法理解通过三种方式/嵌套 has_many :through 关联进行的连接。我有用户与组相关联:通过成员(member)资格。我在多对多
SQL JOIN 和不同类型的 JOIN
什么是 SQL JOIN，有哪些不同的类型？最佳答案插图来自 W3schools : 关于SQL JOIN 和不同类型的 JOIN，我们在Stack Overflow上找到一个类似的问题： htt
Mysql join 使所有 join
我正在尝试访问数据库的两个表。在商店里，我保留了一个事件列表，其中包含 Table Event id, name,datei,houri, dateF,Hourf ,capacity, age ,de
mysql - 复杂连接(Joining Joins)
我有 4 个表:booking、address、search_address 和 search_address_log 表:(相关列) 预订:(pickup_address_id, dropoff_a
Joining after join with yq(在与yq连接之后进行连接)
我在YML中有以下结构：。我正试着创造一个这样的结构：。作业名称和脚本用~分隔，作业用；分隔。。我可以使用以下命令使其正常工作。然而，我想知道是否可以用一个yq表达式来完成，而不是通过管道再次使用yq
Joining after join with yq(在与yq连接之后进行连接)
我在YML中有以下结构：。我正试着创造一个这样的结构：。作业名称和脚本用~分隔，作业用；分隔。。我可以使用以下命令使其正常工作。然而，我想知道是否可以用一个yq表达式来完成，而不是通过管道再次使用yq

行者123

个人简介

我是一名优秀的程序员,十分优秀！

作者热门文章

滴滴打车优惠券免费领取

全站热门文章

首页

博学

6Ren·AI

商城

sql - SQL JOIN 查询中的重复条目