SQL Concat 子查询

Question

我希望“user_names”是一个用户列表，但我无法将多个子查询项分配给“user_names”。如果子查询只返回 1 个项目，它会起作用，但如果它返回多个，则不起作用。

床头柜
[编号]
[1]
[2]
[3]

分配表:
[id, bed_id, user_id]
[1, 1, 1]
[2, 1, 2]

用户表:
[id, '用户名']
[1, '约翰·史密斯']
[2, '李四']

    sql = "SELECT
        b.id,
        (
            SELECT
                u.user_name
            FROM
                assignments AS a
            INNER JOIN
                users as u
            ON
                a.user_id = u.id
            WHERE a.bed_id = b.id
        ) AS user_names
        FROM beds AS b"

期望的结果是:
[1, 'John Smith, Jane Doe']
[2, '']
[3, '']

我尝试对床位 ID 进行硬编码并运行此段以获取姓名列表。它没有用:

    sql = """
        (
        SELECT
            array_agg(user_name)
        FROM
            roomchoice_assignment AS a
        INNER JOIN
            roomchoice_customuser as u
        ON
            a.user_id = u.id
        WHERE
            a.bed_id = 1
        GROUP BY user_name
        )"""

它返回了以下内容:

[
    [
        [
            "John Smith"
        ]
    ], 
    [
        [
            "Jane Doe"
        ]
    ]
]

我希望这样:

['John Smith, Jane Doe']

Answer 1

您的查询的一个问题是您按应用 array_agg 的列进行分组。如果你删除组，你会得到 "{"John Smith","Jane Doe"}"，但你仍然会缺少 bed id 列，如果你想要所有床的列表即使没有分配，您也应该使用左连接而不是子查询(这对于性能和可读性也应该更好)。

您可以按照重复问题的指示使用 string_agg。

这个查询:

SELECT b.id, string_agg(u.user_name, ', ') users
FROM beds AS b
LEFT JOIN assignment AS a ON a.bed_id = b.id
LEFT JOIN users as u ON a.user_id = u.id
GROUP by b.id;

会给你这样的结果:

1;"John Smith, Jane Doe"
2;""
3;""