i want to use inheritance in my database (PostgreSQL) so i wrote this code but finally i found there are inherits in PostgreSQL
CREATE SEQUENCE public.user_account_id_seq;
CREATE TABLE public.user_account (
id BIGINT NOT NULL DEFAULT nextval('public.user_account_id_seq'),
login VARCHAR(50) NOT NULL,
f_name VARCHAR(50),
l_name VARCHAR(50),
email VARCHAR(50) ,
CONSTRAINT user_account_pk PRIMARY KEY (id)
);
CREATE SEQUENCE public.simple_user_account_id_seq;
CREATE TABLE public.simple_user_account (
id BIGINT NOT NULL DEFAULT nextval('public.simple_user_account_id_seq'),
CONSTRAINT simple_user_account_pk PRIMARY KEY (id)
);
CREATE SEQUENCE public.supervisor_account_id_seq;
CREATE TABLE public.supervisor_account (
id BIGINT NOT NULL DEFAULT nextval('public.supervisor_account_id_seq'),
CONSTRAINT supervisor_account_pk PRIMARY KEY (id)
);
with
ALTER TABLE public.simple_user_account ADD CONSTRAINT simple_user_account_fk
FOREIGN KEY (id)
REFERENCES public.user_account (id)
ON DELETE CASCADE
ON UPDATE NO ACTION
NOT DEFERRABLE;
ALTER TABLE public.supervisor_account ADD CONSTRAINT supervisor_account_fk
FOREIGN KEY (id)
REFERENCES public.user_account (id)
ON DELETE CASCADE
ON UPDATE NO ACTION
NOT DEFERRABLE;
or i use INHERITS of postgreSQL
With inheritance you cannot have globally unique constraints. If that doesn't bother you, you can use the feature.
Related
I have three tables that are linked together
My script:
-- Ticket --
CREATE TABLE public.ticket (
id bigint NOT NULL,
libelle character varying(255) NOT NULL,
description character varying(255) NOT NULL,
status character varying(255) NOT NULL,
date_creation timestamp NOT NULL,
date_modification timestamp NOT NULL,
user_createur_id bigint,
referent_realisateur_id bigint,
CONSTRAINT pk_ticket PRIMARY KEY (id)
);
-- Ticket_Avance TABLE --
CREATE TABLE public.ticket_avance (
id bigint NOT NULL,
date_livraison_souhaite timestamp NOT NULL,
date_engagement_livraison timestamp NOT NULL,
referent_demandeur_id bigint
);
ALTER TABLE public.ticket_avance ADD CONSTRAINT "fk_ticket_ticketAvance" FOREIGN KEY (id)
REFERENCES public.ticket (id) MATCH SIMPLE
ON DELETE NO ACTION ON UPDATE NO ACTION;
-- Demande_Travaux TABLE --
DROP TABLE IF EXISTS public.demande_travaux CASCADE;
CREATE TABLE public.demande_travaux (
id bigint NOT NULL,
contrat_id bigint
);
ALTER TABLE public.demande_travaux ADD CONSTRAINT "fk_ticketAvance_DDT" FOREIGN KEY (id)
REFERENCES public.ticket_avance (id) MATCH FULL
ON DELETE NO ACTION ON UPDATE NO ACTION;
I have this error on the demand_travaux creation
SQL Error [42830]: ERROR: there is no unique constraint matching given keys for referenced table "ticket_avance"
ERROR: there is no unique constraint matching given keys for referenced table "ticket_avance"
ERROR: there is no unique constraint matching given keys for referenced table "ticket_avance"
You have forgotten to declare column ID in table ticket_avance as primary key.
Please use the following SQL:
-- Ticket_Avance TABLE --
CREATE TABLE public.ticket_avance (
id bigint NOT NULL,
date_livraison_souhaite timestamp NOT NULL,
date_engagement_livraison timestamp NOT NULL,
referent_demandeur_id bigint,
CONSTRAINT pk_avance PRIMARY KEY (id) --- add this
);
you need add constraint in Ticket_Avance table because you provide reference this id to demande_travaux
-- Ticket_Avance TABLE --
CREATE TABLE ticket_avance (
id bigint NOT NULL,
date_livraison_souhaite timestamp NOT NULL,
date_engagement_livraison timestamp NOT NULL,
referent_demandeur_id bigint,
CONSTRAINT pk_ticket_avance PRIMARY KEY (id) //constraint that you need
);
ALTER TABLE demande_travaux ADD CONSTRAINT "fk_ticketAvance_DDT" FOREIGN KEY (id)
REFERENCES ticket_avance (id) MATCH FULL
ON DELETE NO ACTION ON UPDATE NO ACTION;
here is the demo link of your full query
A foreign key constraint has to target a primary key or unique constraint. The database has to be able to identify a single row in the "parent" table.
You could add primary key constraints:
ALTER TABLE public.ticket_avance ADD PRIMARY KEY (id);
In addition, you should have an index on the column on which the foreign key is defined, particularly if you plan to delete parent rows. With the primary key above, you have such an index on id, but you also should have one on demande_travaux.
The simplest way is to define id as primary key there too:
ALTER TABLE public.demande_travaux ADD PRIMARY KEY (id);
I built a big database with many tables. After that I forgat to add "delete on cascade". How do I add the "delete on casade" on the specific columns I want ?
For example:
create table Users
(
ID char(9) primary key check (ID like replicate('[0-9]',9)),
firstName nvarchar(20) not null,
)
And
create table Applications
(
name nvarchar(20) primary key,
establishDate date not null
)
How do I alter this table-registered to, so that when I delete an app or users it deletes the rows that has this value?
create table RegisteredTo
(
userID char(9) references Users(ID),
ApplicationName nvarchar(20) references Applications(name),
primary key(userID, ApplicationName)
)
I tried something like that - but I got errors.
alter table RegisteredTo
ALTER COLUMN applicationName references Applications(name) ON DELETE CASCADE
nvarchar(20)
You need to add a FOREIGN CONSTRAINT and ON DELETE CASCADE:
ALTER TABLE RegisteredTo
ADD CONSTRAINT FK_RegisteredTo_Users FOREIGN KEY (userID) REFERENCES Users(ID) ON DELETE CASCADE;
ALTER TABLE RegisteredTo
ADD CONSTRAINT FK_RegisteredTo_Applications FOREIGN KEY (ApplicationName) REFERENCES Applications(name) ON DELETE CASCADE;
I have a situation where I want to create a table that associates records from other tables by the id. A constraint of the association is that the year must be the same in the record being associated in each table... Is there a way to get PostgreSQL to CHECK this condition on INSERT?
Table 1:
CREATE TABLE "tenant"."report" (
"id" UUID NOT NULL DEFAULT "pascal".uuid_generate_v1(),
CONSTRAINT "report_pkc_id" PRIMARY KEY ("id"),
"reporting_period" integer NOT NULL,
"name" VARCHAR(64) NOT NULL,
CONSTRAINT "report_uc__name" UNIQUE ("reporting_period", "name"),
"description" VARCHAR(2048) NOT NULL
);
Table 2:
CREATE TABLE "tenant"."upload_file" (
"id" UUID NOT NULL DEFAULT "pascal".uuid_generate_v1(),
CONSTRAINT "upload_file_pkc_id" PRIMARY KEY ("id"),
"file_name" VARCHAR(256) NOT NULL,
"reporting_period" integer
)
Association Table:
CREATE TABLE "tenant"."report_upload_files"
(
"report_id" UUID NOT NULL,
CONSTRAINT "report_upload_files_pkc_tenant_id" PRIMARY KEY ("report_id"),
CONSTRAINT "report_upload_files_fkc_tenant_id" FOREIGN KEY ("report_id")
REFERENCES "tenant"."report" ("id") MATCH SIMPLE
ON UPDATE CASCADE ON DELETE CASCADE,
"upload_file_id" UUID NOT NULL,
CONSTRAINT "report_upload_files_fkc_layout_id" FOREIGN KEY ("upload_file_id")
REFERENCES "tenant"."upload_file" ("id") MATCH SIMPLE
ON UPDATE CASCADE ON DELETE CASCADE
)
I want to add something like to the association table CREATE statement:
CHECK ("tenant"."report"."reporting_period" = "tenant"."upload_file"."reporting_period")
You're solving problems that you've created yourself.
Your data model is a typical one-to-many relationship. You don't need an association table. Also, you don't need the same column in two related tables, one of them is redundant. Use the model as shown below to avoid typical problems resulting from lack of normalization.
create table tenant.report (
id uuid primary key default pascal.uuid_generate_v1(),
reporting_period integer not null,
name varchar(64) not null,
description varchar(2048) not null,
unique (reporting_period, name)
);
create table tenant.upload_file (
id uuid primary key default pascal.uuid_generate_v1(),
report_id uuid references tenant.report(id),
file_name varchar(256) not null
);
Using this approach there's no need to ensure that the reporting periods match between the associated records.
BTW, I would use text instead of varchar(n) and integer (serial) instead of uuid.
Using a TRIGGER function I was able to achieve the desired effect:
CREATE FUNCTION "tenant".report_upload_files_create() RETURNS TRIGGER AS
$report_upload_files_create$
BEGIN
IF NOT EXISTS (
SELECT
*
FROM
"tenant"."report",
"tenant"."upload_file"
WHERE
"tenant"."report"."id" = NEW."report_id"
AND
"tenant"."upload_file"."id" = NEW."upload_file_id"
AND
"tenant"."report"."reporting_period" = "tenant"."upload_file"."reporting_period"
)
THEN
RAISE EXCEPTION 'Report and Upload File reporting periods do not match';
END IF;
RETURN NEW;
END
$report_upload_files_create$ LANGUAGE plpgsql;
CREATE TRIGGER "report_upload_files_create" BEFORE INSERT ON "tenant"."report_upload_files"
FOR EACH ROW EXECUTE PROCEDURE "tenant".report_upload_files_create();
I have a problem with something in SQL, let's see an example of database :
CREATE TABLE person( //Employee
pe_id PRIMARY KEY NOT NULL AUTO_INCREMENT,
pe_name VARCHAR(20),
pe_office VARCHAR(20)
);
CREATE TABLE project( //Mission
pr_id PRIMARY KEY NOT NULL AUTO_INCREMENT,
pr_name VARCHAR(20),
pr_status VARCHAR(15)
);
CREATE TABLE techno( //Programming language
te_id PRIMARY KEY NOT NULL AUTO_INCREMENT,
te_name VARCHAR(20)
);
CREATE TABLE job( //developer, manager, ...
jo_id PRIMARY KEY NOT NULL AUTO_INCREMENT,
jo_name VARCHAR(20)
);
I would like to assign persons on projects for a job using technos.
For example, Rob works as a developer and project manager on the projet #13 with AngularJS and HTML.
So I created this table :
CREATE TABLE assignment(
pe_id INT,
pr_id INT,
te_id INT,
jo_id INT,
as_days INT, //Days of work
PRIMARY KEY(pe_id, pr_id, tr_id, jo_id),
CONSTRAINT fk_as_pe_id FOREIGN KEY(pe_id) REFERENCES person(pe_id),
CONSTRAINT fk_as_pr_id FOREIGN KEY(pr_id) REFERENCES project(pr_id),
CONSTRAINT fk_as_te_id FOREIGN KEY(te_id) REFERENCES techno(te_id),
CONSTRAINT fk_as_jo_id FOREIGN KEY(jo_id) REFERENCES job(jo_id)
);
I would like to have the ability to assign a developer with somes technos to a project without knowing who exaclty, like this:
INSERT INTO assignment(pr_id,te_id,jo_id,as_days) VALUES(1,2,3,4); //No person!
We suppose that this values exists in project, techno and job tables.
But it seems that I can not insert this, probably because I do not define person's ID (which is in the primary key).
How can I do this ?
Hope I'm understandable :)
You solve this problem by not having this as a primary key. Primary keys cannot be NULL or, if they're composite primary keys, cannot contain NULL. Make it a unique index instead. Create an autonumber field for the primary key. I think this is better solution in your case
Primary Key:
Can be only one in a table
It never allows null values
Primary Key is unique key identifier and can not be null and must be unique.
Unique Key:
Can be more than one unique key in one table.
Unique key can have null values(only single null is allowed).
It can be a candidate key
Unique key can be null and may not be unique.
Maybe you should do this:
Before insert disable constraint:
ALTER INDEX fk_as_pe_id ON assignment
DISABLE;
After insert enable it:
ALTER INDEX fk_as_pe_id ON assignment
REBUILD;
Another alternate way is, if it is possible to alter table structure, just exclude pe_id from the composite primary key in assignment table
I'm not sure why these have to be unique, but from reading the MySQL forums it appears that they do. However, I think it has something more to do with the INDEX name. I have two tables that have foreign key constraints referencing the same primary key on a third table. If it helps, I'm using MySQL workbench to design the schema.
I usually name my foreign key on each table the same name as the primary key it references. I guess this isn't possible. It will create the first table with the foreign key constraint, but when it tries to create the second table it throws an error. Here is the second table it throws the error on:
CREATE TABLE IF NOT EXISTS `joe`.`products_to_categories` (
`product_to_category_id` INT NOT NULL AUTO_INCREMENT ,
`category_id` INT NOT NULL ,
`product_id` INT NOT NULL ,
PRIMARY KEY (`product_to_category_id`) ,
INDEX `category_id` (`category_id` ASC) ,
INDEX `product_id` (`product_id` ASC) ,
CONSTRAINT `category_id`
FOREIGN KEY (`category_id` )
REFERENCES `joe`.`categories` (`category_id` )
ON DELETE CASCADE
ON UPDATE NO ACTION,
CONSTRAINT `product_id`
FOREIGN KEY (`product_id` )
REFERENCES `joe`.`products` (`product_id` )
ON DELETE CASCADE
ON UPDATE NO ACTION)
ENGINE = InnoDB;
I want the foreign key names to be the same as the primary key in both of the other tables. What should I remove here so that I can use these names. What is the best practice here.
It is not possible because you would have a conflict in the filename for the file that is used for the index IIRC. I probably would name the key < tablename >_< column_name > or something like that.
You are creating an index (constraint) by the name of product_id via:
INDEX product_id
Then you are going and creating another constraint (for the foreign key) with the same name:
CONSTRAINT product_id
What you need to do is allow the server to provide a default, unique constraint name by removing the
CONSTRAINT product_id
See this URL: http://dev.mysql.com/doc/refman/5.1/en/innodb-foreign-key-constraints.html
"If the CONSTRAINT symbol clause is given, the symbol value must be unique in the database. If the clause is not given, InnoDB creates the name automatically."
In PostgreSQL, the default for naming indexes is to append "_pkey" and "_fkey" to the name of the primary and foreign key, respectively. So your case would look like:
INDEX `product_id_fkey` (`product_id` ASC) ,
UPDATE: I just tried this and it worked. See if that's what you had in mind.
use test;
create table if not exists test.product
(
product_id int not null auto_increment,
name varchar(80) not null,
primary key(product_id)
);
create table if not exists test.category
(
category_id int not null auto_increment,
name varchar(80) not null,
primary key(category_id)
);
create table if not exists test.product_category
(
product_id int,
category_id int,
primary key(product_id, category_id),
constraint product_id_fkey
foreign key(product_id) references product(product_id)
on delete cascade
on update no action,
constraint category_id_fkey
foreign key(category_id) references category(category_id)
on delete cascade
on update no action
);
insert into test.product(name) values('teddy bear');
insert into test.category(name) values('toy');
insert into test.product_category
select p.product_id, c.category_id from product as p, category as c
where p.name = 'teddy bear' and c.name = 'toy';