我有一个 PostgreSQL 数据库，我使用 node.js 和 pg 模块从中查询数据。我的 Javascript 代码(在服务器端)中的结果数组包含大量重复值，因为这就是返回数据的方式。我相信查询没问题，因此值是重复的，因为多个 INNER JOINS 的表彼此独立，因此 A 和 的结果B 来自一个表(本例中为 qa_layers)，结果为 C 和 D 来自另一表(qa_cases 在本例中)，以及来自第三个表的公共(public)值 E (在本例中为 qa_settings)将表示为
| settings_name | layer_name | case_name |
| E | A | C |
| E | A | D |
| E | B | C |
| E | B | D |
据我所知，这应该是预期的行为。我需要做的是将每列的唯一数据过滤为具有唯一值的数组，例如[A, B] 和 [C, D] 用于 layer_name 和 case_name。在我看来，可能有多种方法可以实现这一目标，但每种方法似乎都有点痛苦，除非有一些我不知道的帮助程序库或功能。
- 用 Javascript 解析数据库中的结果数组。由于存在多个 for 循环，这看起来很痛苦，但如果有一些库可以满足我的需要，这可能会很容易。
- 增强
PostgreSQL查询，以便不需要或更少的解析。我不知道这是否可行，但如果生成的数组中有更容易解析的对象，则解析可能会不那么痛苦。我的查询粘贴在最后。 - 对数据库执行多次查询以获得正确的数据。对我来说，这似乎是可扩展性最差的解决方案，并且解析数组而不是运行多个连续查询时的开销会少很多。
我的 PostgreSQL 查询:
SELECT run.id, settings.name AS settings_name, layer.name AS layer_name, qa_case.name AS case_name FROM qa_runs AS run
INNER JOIN qa_composites_in_run AS cr
ON cr.run_id = run.id
INNER JOIN qa_layers_in_composite AS lc
ON lc.composite_name = cr.composite_name
INNER JOIN qa_layers AS layer
ON layer.name = lc.layer_name
INNER JOIN qa_cases_in_run AS case_run
ON case_run.run_id = run.id
INNER JOIN qa_cases AS qa_case
ON qa_case.name = case_run.case_name
INNER JOIN qa_settings AS settings
ON settings.name = run.settings_name
WHERE run.id IN (27,28,29);
我实际生成的 Javascript 数组:
[
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitLondon"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitRotterdam"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "wtf"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitLondon"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitRotterdam"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "OpenStreetMapService",
"case_name": "wtf"
},
{
"id": 28,
"settings_name": "QA Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitLondon"
},
{
"id": 28,
"settings_name": "QA Default",
"layer_name": "OpenStreetMapService",
"case_name": "wtf"
},
{
"id": 27,
"settings_name": "QA Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitLondon"
},
{
"id": 27,
"settings_name": "QA Default",
"layer_name": "OpenStreetMapService",
"case_name": "VisitRotterdam"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "VisitLondon"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "VisitRotterdam"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "wtf"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "VisitLondon"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "VisitRotterdam"
},
{
"id": 29,
"settings_name": "Test Default",
"layer_name": "Map2D",
"case_name": "wtf"
},
{
"id": 28,
"settings_name": "QA Default",
"layer_name": "Map2D",
"case_name": "VisitLondon"
},
{
"id": 28,
"settings_name": "QA Default",
"layer_name": "Map2D",
"case_name": "wtf"
},
{
"id": 27,
"settings_name": "QA Default",
"layer_name": "Map2D",
"case_name": "VisitLondon"
},
{
"id": 27,
"settings_name": "QA Default",
"layer_name": "Map2D",
"case_name": "VisitRotterdam"
}
]
我希望数组/JSON最终如何:
[
{
"id": 27,
"settings_name": "QA Default",
"layer_names": [
{
"layer_name": "OpenStreetMapService"
},
{
"layer_name": "Map2D"
}
],
"case_names": [
{
"case_name": "VisitLondon"
},
{
"case_name": "VisitRotterdam"
}
]
},
{
"id": 28,
"settings_name": "QA Default",
"layer_names": [
{
"layer_name": "OpenStreetMapService"
},
{
"layer_name": "Map2D"
}
],
"case_names": [
{
"case_name": "VisitLondon"
},
{
"case_name": "wtf"
}
]
},
{
"id": 29,
"settings_name": "Test Default",
"layer_names": [
{
"layer_name": "OpenStreetMapService"
},
{
"layer_name": "Map2D"
}
],
"case_names": [
{
"case_name": "VisitLondon"
},
{
"case_name": "VisitRotterdam"
},
{
"case_name": "wtf"
}
]
}
]
最佳答案
我建议在 PostgreSQL 端执行此操作，通过使用嵌套数据构造并返回一个 json 对象，而不是返回联接关系。
借助 9.3 及更高版本中的 json 支持，您可以相当轻松地完成此操作。如果没有示例数据和架构，我无法真正为您转换查询，但您需要从 json_agg 和 row_to_json 开始。查看其他相关问题，其答案也引用了这些函数。
这是一个简单的例子:
CREATE TABLE parent(
id integer primary key,
parentdata text
);
CREATE TABLE child(
id integer primary key,
parent_id integer not null references parent(id),
childdata text
);
INSERT INTO parent(id, parentdata) VALUES
(1, 'p1'), (2, 'p2'), (3, 'p3');
INSERT INTO child(id, parent_id, childdata)
VALUES
(10, 1, 'c1_10'),
(20, 2, 'c2_20'),
(21, 2, 'c2_21');
SELECT row_to_json(toplevel, true)
FROM (
SELECT p.id, p.parentdata, json_agg(c) AS children
FROM parent p
LEFT OUTER JOIN child c ON (p.id = c.parent_id) GROUP BY p.id
) toplevel;
它发出:
{"id":1,
"parentdata":"p1",
"child":[{"id":10,"parent_id":1,"childdata":"c1_10"}]}
{"id":2,
"parentdata":"p2",
"child":[{"id":20,"parent_id":2,"childdata":"c2_20"},
{"id":21,"parent_id":2,"childdata":"c2_21"}]}
{"id":3,
"parentdata":"p3",
"children":[null]}
如果您必须消除重复，您可能需要确保 SQL 中的 ORDER BY 子句已完全指定，以便行从外到内排序。目的。然后，您可以在 js 中线性扫描表并忽略对象 id 与前一行中同一实体的 id 相同的行。这是一个非常简单的单个 for 循环，带有一些状态跟踪变量。缺点是 PostgreSQL 必须对结果集进行排序。
关于javascript - 在 Javascript 中从 PostgreSQL 查询结果中过滤掉唯一数据，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/30618181/