I usually work with foreign key for good practices and to maintain data integrity but I always face the following "problem" when trying to delete some data:
I have a scenario with a main group table: groups
And I have several other tables that relate to groups, for example:
user_group (relationship table between users and groups)
course_group (relationship table between courses and groups)
When deleting a group I always need to delete the reference to that group in the related tables before ... but currently I have several relationship tables and my exclusion code ends up getting quite large.
My question is: is there any function or library that I could use so that this work could be done more automatically? A function that may check for me all these references and already exclude. If anyone has any idea how to make it faster and with the smallest possible number of code thanks. Note: I am currently working with PHP
You want on delete cascade as an option in the foreign key constraint in the children table.
Say that table groups looks like:
create table groups (
id int primary key,
... -- other columns here
);
You would then declare child table user_groups as:
create table user_groups (
... -- some columns here
group_id int,
foreign key (group_id) references groups(id) on delete cascade
)
Related
I am trying to make online examination system. I have three tables. Table "student" has details of students, column id identifies students uniquely. Table "tests" has details of all the tests available and there is a column id which contains id of tests. and then I have tables of tests which contains questions and answers.
Now I want that if any student has appeared in a test, it can not reappear in that test again. How can I achieve that? Please help. Thanks in advance.
Aah. If you need to create the relation table... First make a table relating the student_id and the test_id. When you make this you can make the primary key on multiple columns:
CREATE TABLE students_tests
(
FOREIGN KEY student_id REFERENCES student(id) ,
FOREIGN KEY test_id REFERENCES tests(id)
PRIMARY KEY (student_id,test_id)
)
However, if you already have the relation table with a primary key, you can make pairs (or tuples) of keys a unique set with the ALTER TABLE ... ADD UNIQUE command . This question has been answered here:
How do I specify unique constraint for multiple columns in MySQL?
i have three tabs in my Db:
1.pers_info(id(primary), name ,....)
2.contacts(c_id(primary), phone, email, ...)
now 1 person can have multiple rows in contacts tab.
thus to minimize redundancy i made another tab contact_relation(id (foregin key references pers_info(id), c_id (foregin key references contacts(c_id))
i successfully created the relation and also was able to insert apt entries(related id and c_id) to it "using last_insert_id();" to extract id and c_id required.
now the problem..
i dropped tab contacts_rel. and created it this way :
CREATE TABLE contacts_relation (
id INT NOT NULL,
cid INT NOT NULL,
FOREIGN KEY (id)
REFERENCES pers_info(id)
ON UPDATE CASCADE ON DELETE RESTRICT,
FOREIGN KEY (c_id)
REFERENCES contacts(c_id)
);
if now i try to insert into tab cont_rel, it gives error :
cannot add or update a child row foreign key constraint fails.
which makes sense..
i was hoping that adding constraints in the defn of tab cont_rel would save the hassle of inserting entries in it manually using "last_insert_id();"
so is there a way, i could maintain ref integrity with new data
coming.. thanks.
Why the third table? When you many to many relation you need intermediate table. You have one to many relation so two tables are sufficient. Table contacts need FK to table pers_info. Add it if you don't have already.
Here's a simple transaction example with LAST_INSERT_ID(): SQL INSERT INTO multiple tables
Some info about foreign keys and altering table:
Add Foreign Key to existing table
http://dev.mysql.com/doc/refman/5.0/en/alter-table.html
I'm designing a website of book library for my project in php.
Anyways I have three tables named books,users and savebook.
Table book has following columns: "bookid", "title","author" "genre" and "summary".
And table users has following columns userid, username, password, and name.
Users can save books as favorites and those saved books are saved in the table named savebook with columns bookid and userid which are foreign key to table book and users
I used following query for that:
ALTER TABLE savebook
ADD CONSTRAINT bkid_usid
FOREIGN KEY (bookid)
REFERENCES books (bookid);
and
ALTER TABLE savebook
ADD CONSTRAINT usid_bkid
FOREIGN KEY (userid)
REFERENCES users(userid);
Now the problem is whenever i try to delete a book from table book using query
DELETE FROM books
WHERE bookid=1;
I get this message:
1451 - Cannot delete or update a parent row: a foreign key constraint fails (booklibrary.savebook, CONSTRAINT bkid_usid FOREIGN KEY (bookid) REFERENCES books (bookid))
How do i delete a book from table book which also deletes the related row in table savebook?
To get the behavior you describe, you can specify
ON DELETE CASCADE
as part of the foreign key definition.
Reference: http://dev.mysql.com/doc/refman/5.5/en/create-table-foreign-keys.html
If you were to modify your foreign key constraint bkid_usid (on the booksave table), then the delete statement that you show
DELETE FROM books WHERE ...
Would cause MySQL to delete the rows in booksave that have a foreign key values that reference rows being removed from books.
ALTER TABLE savebook
ADD CONSTRAINT bkid_usid
FOREIGN KEY (bookid)
REFERENCES books (bookid)
ON DELETE CASCADE
;
Make note of that last line I added to the ALTER from the original question, it is an optional part of the foreign key definition (there is also ON UPDATE ...). By default, when not specified, I believe MySQL treats it as NO ACTION or RESTRICT (those two are effectively the same as far as I know) instead of CASCADE. Full documentation found at http://dev.mysql.com/doc/refman/5.6/en/create-table-foreign-keys.html
An interesting question :)
I think the problem is caused by the so called "referential integrity".
That means, that you are not able to delete a dataset, which primary key is used in another tables dataset as foreign key. If you could do this, some foreign keys of the table savebook would refer to a dataset in table books, that doesn't exist anymore.
So first you have to delete all datasets in savebook, where the bookid is the same one as of the book you want to delete and after that, you can delete this dataset from books. ;)
You have applied constraints to the tables. You need to tell MySQL to delete referenced entries from "savebook" table if any entry is deleted from "book" table.
ALTER TABLE savebook ADD CONSTRAINT bkid_usid FOREIGN KEY (bookid) REFERENCES books (bookid) ON DELETE CASCADE
I need to store multiple id's in either a field in the table or add another table to store the id's in.
Each member will basically have favourite articles. Each article has an id which is stored when the user clicks on a Add to favourites button.
My question is:
Do I create a field and in this field add the multiple id's or do I create a table to add those id's?
What is the best way to do this?
This is a many-to-many relationship, you need an additional table storing pairs of user_id and article_id (primary keys of user and article tables, respectively).
You should create a new table instead of having comma seperated values in a single column.
Keep your database normalized.
You create a separate table, this is how things work in a relational database. The other solution (comma separated list of ids in one column) will lead to an unmaintainable database. For example, what if you want to know how many times an article was favorited? You cannot write queries on a column like this.
Your table will need to store the user's id and the article's id - these refer to the primary keys of the corresponding tables. For querying, you can either use JOINs or nested SELECT queries.
As lafor already pointed out this is a many-to-many relationship and you'll end up with three tables: user, article, and favorite:
CREATE TABLE user(
id INT NOT NULL,
...
PRIMARY KEY (id)
) ENGINE=INNODB;
CREATE TABLE article (
id INT NOT NULL,
...
PRIMARY KEY (id)
) ENGINE=INNODB;
CREATE TABLE favorite (
userID INT NOT NULL,
articleID INT NOT NULL,
FOREIGN KEY (userID) REFERENCES user(id) ON DELETE CASCADE,
FOREIGN KEY (articleID) REFERENCES article(id) ON DELETE CASCADE,
PRIMARY KEY (userID, articleID)
) ENGINE=INNODB;
If you then want to select all user's favorite articles you use a JOIN:
SELECT * FROM favorite f JOIN article a ON f.articleID = a.id WHERE f.userID = ?
If you want to know why you should use this schema, I recommend reading about database normilization. With multiple IDs in a single field you would even violate the first normal form and thus land in a world of pain...
I'm trying to create some tables in a mysql db to handle customers, assign them to groups and give customers within these groups unique promotion codes/coupons.
there are 3 parent(?) tables - customers, groups, promotions
then I have table - customerGroups to assign each customer_id to many group_id's
also I have - customerPromotions to assign each customer_id to many promotion_id's
I know I need to use cascade on delete and update so that when I delete a customer, promotion or group the data is also removed from the child tables. I put together some php to create the tables easily http://pastebin.com/gxhW1PGL
I've been trying to read up on cascade, foreign key references but I think I learn better by trying to do things then learning why they work. Can anyone please give me their input on what I should do to these tables to have them function correctly.
I would like to have the database and tables set up correctly before I start with queries or anything further so any advice would be great.
You seem to want just a little guidance. So I'll try to be brief.
$sql = "CREATE TABLE customerGroups (
customer_id int(11) NOT NULL,
group_id int(11) NOT NULL,
PRIMARY KEY (customer_id, group_id),
CONSTRAINT customers_customergroups_fk
FOREIGN KEY (customer_id)
REFERENCES customers (customer_id)
ON DELETE CASCADE,
CONSTRAINT groups_customergroups_fk
FOREIGN KEY (group_id)
REFERENCES groups (group_id)
ON DELETE CASCADE
)ENGINE = INNODB;";
You only need id numbers when identity is hard to nail down. When you're dealing with people, identity is hard to nail down. There are lots of people named "John Smith".
But you're dealing with two things that have already been identified. (And identified with id numbers, of all things.)
Cascading deletes makes sense. It's relatively rare to cascade updates on id numbers; they're presumed to never change. (The main reason Oracle DBAs insist that primary keys must always be ID numbers, and that they must never change is because Oracle can't cascade updates.) If, later, some id numbers need to change for whatever reason, you can alter the table to include ON UPDATE CASCADE.
$sql = "CREATE TABLE groups
(
group_id int(11) NOT NULL AUTO_INCREMENT,
group_title varchar(50) NOT NULL UNIQUE,
group_desc varchar(140),
PRIMARY KEY (group_id)
)ENGINE = INNODB;";
Note the additional unique constraint on group_title. You don't want to allow anything like this (below) in your database.
group_id group_title
--
1 First group
2 First group
3 First group
...
9384 First group
You'll want to carry those kinds of changes through all your tables. (Except, perhaps, your table of customers.)