Question

当我在某些表之间进行内部联接时，我遇到了一些抛出欺骗的问题。

基本上，我正在加入一些基于videoId的标签和频道（思考类型），因此特定的视频文件可以包含多个标签并成为多个频道的一部分。

不幸的是，出现了多个匹配:(我的直觉是，查询与每个连接的videoId匹配并吐出双倍（正如您将在链接的小提琴中看到的那样，但考虑到我的SQL加入的总数）经验是从大约3个小时前到目前为止，我已经设法混乱，但现在可以做一些专家建议，从这里去哪里。

以下是我一直在努力解决的架构和查询，这里是sqlfiddle链接，

架构：

create table videos 
(
   videoId int(1) AUTO_INCREMENT,
   videoUUID char(40),
   contentFolder varchar(50), 
   title varchar(50), 
   caption varchar(50),
   duration varchar(15),
   date TIMESTAMP,
   url varchar(55),
   text varchar(250), 
   PRIMARY KEY (videoId)
);

create table tags 
( 
   tagId int(1) AUTO_INCREMENT,
   tagName varchar(15),
   PRIMARY KEY (tagId)
);

create table channels
(
   channelId int(1) AUTO_INCREMENT,
   channelName varchar(15),
   PRIMARY KEY (channelId)
);

create table videoTags 
(
   videoTagId int(1) AUTO_INCREMENT,
   videoId int(1),
   tagId int(1),
   PRIMARY KEY (videoTagId)
);

create table videoChannels 
(
    videoChannelId int(1) AUTO_INCREMENT,
    videoId int (1),
    channelId int (1),
    PRIMARY KEY (videoChannelId)
 );

//
create trigger tuuid before insert on videos
for each row begin
set new.videoUUID = uuid();
end//

insert into videos 
  (contentFolder, title,caption,duration,url,text)
values ("someDir/","A Movie Title", "Headline for Movie",
    "00:05:11",
    "http://someserver.ip/somedir/test.mp4",
    "Some text as part of the video file description here");

insert into tags (tagName) values
('Flowers'),('Dogs'),('Cats'),('YaMum'),('orlyowl');

insert into channels (channelName) values
('General'), ('NotSoGeneral'), ('Specific'), ('Broad'), ('Narrow'),
('Obsolete'); 

insert into videoTags (videoId,tagId) values
(1,2),(1,5);

insert into videoChannels (videoId,channelId) values
(1,1),(1,4),(1,6);

查询：

select distinct v.*,group_concat(t.tagName)Tags,
group_concat(c.channelName)Channels
from videos as v 

inner join videoTags as vt on v.videoId = vt.videoid

inner join tags as t on t.tagId = vt.tagId

inner join videoChannels as vc on v.videoId = vc.videoId

inner join channels as c on c.channelId = vc.channelId

group by v.videoId;

我是以正确的方式来做这件事的吗？我构建查询的方式是否存在根本的错误？或者说我想要完成的模式是错误的。

非常感谢任何帮助！

@J Lo谢谢你的回答！

答案： DISTINCT关键字，我在select语句中尝试过，但不确定它的用法，所以没有完全理解它可以在哪里使用，@ J Lo如此简洁地指出，我可以在我的group_concat中使用它，从而给出了查询：

select v.*,group_concat(distinct t.tagName)Tags,
group_concat(distinct c.channelName)Channels
from videos as v 

inner join videoTags as vt on v.videoId = vt.videoid

inner join tags as t on t.tagId = vt.tagId

inner join videoChannels as vc on v.videoId = vc.videoId

inner join channels as c on c.channelId = vc.channelId

group by v.videoId;

Answer 1

你的第一对夫妇加入（视频/视频标签/标签）产生如下表格：

VideoID = 1 will bring in TagID = 2,5 (Dogs, orlyowl) so you have this

| 1 | Dogs
| 1 | orlyowl

当您加入VideoChannels时，它会复制每个频道的上述条目

| 1 | Dogs    | 1
| 1 | orlyowl | 1
| 1 | Dogs    | 4
| 1 | orlyowl | 4
| 1 | Dogs    | 6
| 1 | orlyowl | 6

group_concat具有DISTINCT属性

select v.*
  , group_concat(distinct t.tagName) Tags
  , group_concat(distinct c.channelName) Channels
from videos as v 
inner join videoTags as vt on v.videoId = vt.videoid
inner join tags as t on t.tagId = vt.tagId
inner join videoChannels as vc on v.videoId = vc.videoId
inner join channels as c on c.channelId = vc.channelId
group by v.videoId;

SQL多表连接投掷欺骗

1 个答案: