为什么我不能对递归视图发出与对支持它的表相同的查询呢?

qojgxg4l  于 2021-08-01  发布在  Java
关注(0)|答案(1)|浏览(359)

我有一个 tags 表,我在它上面定义了一个递归cte视图,其中包含tags表的所有列:

CREATE VIEW tags_paths AS
WITH RECURSIVE tag_path (id, created_at, updated_at, community_id, tag_set_id, wiki_markdown,
                         wiki, excerpt, parent_id, name, path) AS
                   (
                       SELECT id, created_at, updated_at, community_id, tag_set_id, wiki_markdown,
                              wiki, excerpt, parent_id, name, name as path
                       FROM tags
                       WHERE parent_id IS NULL
                       UNION ALL
                       SELECT t.id, t.created_at, t.updated_at, t.community_id, t.tag_set_id,
                              t.wiki_markdown, t.wiki, t.excerpt, t.parent_id, t.name,
                              concat(tp.name, ' > ', t.name) as path
                       FROM tag_path AS tp JOIN tags AS t ON tp.id = t.parent_id
                   )
SELECT * FROM tag_path
ORDER BY path;

我可以针对 tags table很好:

SELECT  tags.*, COUNT(posts.id) AS post_count
FROM `tags`
    LEFT OUTER JOIN `posts_tags` ON `posts_tags`.`tag_id` = `tags`.`id`
    LEFT OUTER JOIN `posts` ON `posts`.`community_id` = 2 AND `posts`.`id` = `posts_tags`.`post_id`
WHERE `tags`.`community_id` = 2 AND `tags`.`tag_set_id` = 3
GROUP BY tags.id ORDER BY COUNT(posts.id) DESC LIMIT 96 OFFSET 0;

但是,针对 tags_paths 查看:

SELECT  tags_paths.*, COUNT(posts.id) AS post_count
FROM `tags_paths`
    LEFT OUTER JOIN `posts_tags` ON `posts_tags`.`tag_id` = `tags_paths`.`id`
    LEFT OUTER JOIN `posts` ON `posts`.`community_id` = 2 AND `posts`.`id` = `posts_tags`.`post_id`
WHERE `tags_paths`.`community_id` = 2 AND `tags_paths`.`tag_set_id` = 3
GROUP BY tags_paths.id ORDER BY COUNT(posts.id) DESC LIMIT 96 OFFSET 0;

作为错误返回,特别是:

[42000][1055] Expression #2 of SELECT list is not in GROUP BY clause and contains nonaggregated column
'tags_paths.created_at' which is not functionally dependent on columns in GROUP BY clause; this is
incompatible with sql_mode=only_full_group_by

为什么?我能怎么办?

noj0wjuj

noj0wjuj1#

作为@秋叶树州 id table上有把钥匙吗 tags ,但不在视图上。因此,不能在视图中为其余列建立与 id 列,由sql标准指定。
你有两个选择。您可以:
将所有列添加到 GROUP BY 条款。有时,对于大型类型(如 TEXT 或者二进制文件。
或者,可以聚合 SELECT 条款。
下面是使用后者修改的sql语句:

SELECT
  tags_paths.id,
  max(tags_paths.created_at) as created_at,   
  max(tags_paths.updated_at) as updated_at, 
  max(tags_paths.community_id) as community_id, 
  max(tags_paths.tag_set_id) as tag_set_id, 
  max(tags_paths.wiki_markdown) as markdown,
  max(tags_paths.wiki) as wiki, 
  max(tags_paths.excerpt) as excerpt, 
  max(tags_paths.parent_id) as parent_id, 
  max(tags_paths.name) as name, 
  max(tags_paths.path) as path,
  COUNT(posts.id) AS post_count
FROM `tags_paths`
LEFT OUTER JOIN `posts_tags` ON `posts_tags`.`tag_id` = `tags_paths`.`id`
LEFT OUTER JOIN `posts` ON `posts`.`community_id` = 2 
                       AND `posts`.`id` = `posts_tags`.`post_id`
WHERE `tags_paths`.`community_id` = 2 AND `tags_paths`.`tag_set_id` = 3
GROUP BY tags_paths.id 
ORDER BY COUNT(posts.id) DESC 
LIMIT 96 
OFFSET 0;

是的,它变长了。

相关问题