Postgres，无法在json_aggr中使用distinct

Question

我不太擅长SQL，我有三个表机构、模块、文档。模块和文档属于公司。 我想选择一家拥有所有信息并按状态汇总的公司：

select "agencies".*,
JSON_AGG(  modules.*) FILTER(WHERE modules.status like 'active') as activeModules,
JSON_AGG(  modules.*) FILTER(WHERE modules.status like 'draft') as draftModules,
JSON_AGG(  documents.*) FILTER(WHERE documents.status like 'active') as activeDocuments
from "agencies"
left outer join "modules" on "agencies"."id" = "modules"."agency_id"
left outer join "documents" on "agencies"."id" = "documents"."agency_id"
where "agencies"."id" = '959e5e04-e8ba-4367-b3d7-8bc5d5b5e666' group by "agencies"."id"

这可行，但聚合的所有记录都是重复的

如果我像这样更改选择：

JSON_AGG(  DISTINCT modules.*) FILTER(WHERE modules.status like 'active') as activeModules,

它引发错误：错误：无法识别类型模块的相等运算符

我通过这样做找到了解决方案：

JSON_AGG( distinct jsonb_build_object(modules.id, modules.field2...) ) FILTER(WHERE modules.status like 'active') as activeModules,

这可行，但我不喜欢指定所有字段的想法。 有没有一种方法可以实现我想要的而不需要插入所有字段？

Answer 1

您可以尝试编写查询以使用子查询来查找 JSON 聚合：

SELECT a.*, m.activeModules, m.draftModules, d.activeDocuments
FROM agencies a
LEFT JOIN
(
    SELECT agency_id,
           JSON_AGG(*) FILTER (WHERE status LIKE 'active') AS activeModules,
           JSON_AGG(*) FILTER (WHERE status LIKE 'draft') AS draftModules
    FROM modules
    GROUP BY agency_id
) m
    ON a.id = m.agency_id
LEFT JOIN
(
    SELECT agency_id,
           JSON_AGG(*) FILTER (WHERE status LIKE 'active') AS activeDocuments
    FROM documents
    GROUP BY agency_id
) d
    ON a.id = d.agency_id
WHERE a.id = '959e5e04-e8ba-4367-b3d7-8bc5d5b5e666'
GROUP BY a.id;