Welcome toVigges Developer Community-Open, Learning,Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
1.6k views
in Technique[技术] by (71.8m points)

mysql - sql join tables where 1 column has comma

how can I get all usernames when I search "new1" .For eg: I should get A and B as userids 1,2 in tblC is 1,2 for row1 which has new1.What query should I use to get the above result? I really appreciate any help.Thanks in Advance.
http://sqlfiddle.com/#!2/1ab8e/2

CREATE TABLE if not exists tblA
(
id int(11) NOT NULL auto_increment ,
user varchar(255),
 category int(255),
 PRIMARY KEY (id)
);

CREATE TABLE if not exists tblB
(
id int(11) NOT NULL auto_increment ,
username varchar(255),
 userid int(255),
 PRIMARY KEY (id)
);

CREATE TABLE if not exists tblC
(
id int(11) NOT NULL auto_increment ,
nname varchar(255),
 userids varchar(255),
 PRIMARY KEY (id)
);


INSERT INTO tblA (user, category ) VALUES
('1', '1'),
('1', '2'),
('1', '3'),
('1', '1'),
('2', '1'),
('2', '1'),
('2', '1'),
('2', '1'),
('3', '1'),
('2', '1'),
('4', '1'),
('4', '1'),
('2', '1');


INSERT INTO tblB (userid, username ) VALUES
('1', 'A'),
('2', 'B'),
('3', 'C'),
('4', 'D'),
('5', 'E');


INSERT INTO tblC (id, nname,userids ) VALUES
('1', 'new1','1,2'),
('2', 'new2','1,3'),
('3', 'new3','1,4'),
('4', 'new4','3,2'),
('5', 'new5','5,2');

Query so far:

select * where nname="new1" from  tblC
CROSS JOIN tblB
ON tblB.userid=(SELECT userids FROM substr(tblC.userids,','))
See Question&Answers more detail:os

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Answer

0 votes
by (71.8m points)

You should really look at Database normalization and first normalize your structure by adding a junction table and holds a relation from tablec each relation stored in tablec will be stored in new junction table but not as comma separated list each row will hold id of c and one user id per row ,if you can't alter your schema you can use find_in_set to find values in set

select *  
from  tblC c
JOIN tblB b
ON (find_in_set(b.userid,c.userids) > 0)
where c.nname="new1"

See demo


Edit for normalize schema

I have removed userids column from your tblC and instead i have created a new junction table as tblC_user with 2 columns c_id this will related to the id column of tblC and second one userid to store user relations users for tblC see sample schema for tblC

CREATE TABLE if not exists tblC
(
id int(11) NOT NULL auto_increment ,
nname varchar(255),
 PRIMARY KEY (id)
);

INSERT INTO tblC (id, nname) VALUES
('1', 'new1'),
('2', 'new2'),
('3', 'new3'),
('4', 'new4'),
('5', 'new5');

And here is your junction table as tblC_user

CREATE TABLE if not exists tblC_user
(
 c_id int,
 userid int
);

INSERT INTO tblC_user (c_id,userid) VALUES
('1','1'),
('1','2'),
('2','1'),
('2','3'),
('3','1'),
('3','4'),
('4','3'),
('4','2'),
('5','5'),
('5','2');

In above if you notice i haven't stored any comma separated relations each relation of user for tblC is stored in new row ,for you concerned result set i have used junction table in join also new query will be like below

select *  
from  tblC c
join tblC_user cu on(c.id = cu.c_id)
join tblB b on (b.userid = cu.userid)
where c.nname="new1"

Demo 2

Now above query can can be optimized by using indexes you can maintain cascading relations easily


与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome to Vigges Developer Community for programmer and developer-Open, Learning and Share
...