I have a PostgreSQL database, 3 tables and my schema as follows.
CREATE TABLE table_a (
id SERIAL PRIMARY KEY,
name_a VARCHAR(255)
);
CREATE TABLE table_b (
name_b VARCHAR(255)
) INHERITS (table_a);
CREATE TABLE table_c (
name_c VARCHAR(255)
) INHERITS (table_a);
insert into table_b (name_a, name_b) values('table A','table B1');
insert into table_b (name_a, name_b) values('table A','table B2');
insert into table_c (name_a, name_c) values('table A','table C1');
insert into table_c (name_a, name_c) values('table A','table C2');
select * from table_a;
select * from table_b;
select * from table_c;
Now i want to add an association between Table Band Table C like this:
I do not know if this is possible when we inherit the same table ?
I do not see how I can create this association ?
You need a unique identifier or primary key on table_b. Quoting docs:
All check constraints and not-null constraints on a parent table are automatically inherited by its children, unless explicitly specified otherwise with NO INHERIT clauses. Other types of constraints (unique, primary key, and foreign key constraints) are not inherited.
CREATE TABLE table_a (
id SERIAL PRIMARY KEY,
name_a VARCHAR(255)
);
CREATE TABLE table_b (
name_b VARCHAR(255),
primary key (id) --> here you set the PK
) INHERITS (table_a);
CREATE TABLE table_c (
name_c VARCHAR(255),
id_b int references table_b(id) --> the fk to b through pk
) INHERITS (table_a);
Alternative way
In your second diagram, do you have a kind of identifier for table_b. It is right because you can reference table by a unique field. In this case, the DDL will be like this:
CREATE TABLE table_a (
id SERIAL PRIMARY KEY,
name_a VARCHAR(255)
);
CREATE TABLE table_b (
name_b VARCHAR(255),
id_b SERIAL UNIQUE --> Here the unique id for b
--, primary key (id) -- optionally
) INHERITS (table_a);
CREATE TABLE table_c (
name_c VARCHAR(255),
id_b int references table_b(id_b) --> the fk to b through unique
) INHERITS (table_a);
I prefer the first approach, but I have posted also this one just for academical purposes.
Related
I have the following tables: (notice that SUPP may be NULL)
CREATE TABLE IF NOT EXISTS A
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS B
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS C
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS D
(
SID INTEGER NOT NULL,
DID INTEGER NOT NULL,
SUPP INTEGER,
PRIMARY KEY (SID, DID),
FOREIGN KEY (SID) REFERENCES A (ID),
FOREIGN KEY (DID) REFERENCES B (ID),
FOREIGN KEY (SUPP) REFERENCES C (ID)
);
Assuming table A contains ID=2 and table B contains ID=5.
I want to perform the following insert:
INSERT OR REPLACE INTO D VALUES (2,5,NULL);
However I get the following error:
Foreign key mismatch - D referencing C (INSERT INTO "main"."D" ...
I'm assuming the issue stems from the NULL foreign key but I don't know what I did wrong or how to fix it.
You need to use valid referenced value if you are inserting into child table. The ID 2 in table A and ID 5 in table B has to be present.
CREATE TABLE IF NOT EXISTS A
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS B
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS C
(
ID INTEGER PRIMARY KEY NOT NULL
);
CREATE TABLE IF NOT EXISTS D
(
SID INTEGER NOT NULL,
DID INTEGER NOT NULL,
SUPP INTEGER,
PRIMARY KEY (SID, DID),
FOREIGN KEY (SID) REFERENCES A (ID),
FOREIGN KEY (DID) REFERENCES B (ID),
FOREIGN KEY (SUPP) REFERENCES C (ID)
);
INSERT INTO A VALUES (2);
INSERT INTO B VALUES (5);
INSERT INTO D VALUES (2, 5, NULL);
First of all I´m working with DB2.
My problem:
I have a table A with a primary key.
A table B with 2 primary keys (one of them is a foreign key of A)
A table C with a primary key
A table D which has got the primary keys of B and C
How can I create table D?
My first idea was that, but it doesn't work:
Create Table D(
A varchar(20) not null references B(A),
B varchar(20) not null references B(name of prim key attribute from B),
C varchar(20) not null references C,
primary key(A,B,C)
);
Hope you understand my problem and can help me.
Foreign keys reference the names of columns, not the names of constraints.
The set up . . .
create table A (
col_a int primary key
);
create table B (
col_a int not null,
col_b int not null,
primary key (col_a, col_b),
foreign key (col_a) references A (col_a)
);
create table C (
col_c int primary key
);
And the execution . . .
create table D (
col_a int not null,
col_b int not null,
col_c int not null,
primary key (col_a, col_b, col_c),
foreign key (col_a, col_b) references B (col_a, col_b),
foreign key (col_c) references C (col_c)
);
As Marc_s commented there can only be ONE primary key in a table ie, you cannot have more than one primary key in a table.
As a solution to your problem you can probably combine two columns of table B and table C which can together act as a primary key for table D
I've got three tables: A, B and C. I want to create table A, and it should have 3 columns, first one should be a PRIMARY KEY.
1) How to create second column, with is a refference to table B,
2) and third with is a refference to 'C.id' row from table C. A.id = C.id
CREATE TABLE A
(
id SERIAL PRIMARY KEY,
// ? - reference to table B
// ? - reference to C.id row. A.id = C.id
)
Database: postgresql
I assume you use mysql database.
CREATE TABLE A
(
id INT NOT NULL PRIMARY KEY,
b_id INT NOT NULL,
c_id INT NOT NULL,
FOREIGN KEY (b_id) REFERENCES B (id),
FOREIGN KEY (c_id) REFERENCES C (id)
) TYPE = INNODB;
Update for using postgresql:
CREATE TABLE "A"
(
id integer NOT NULL,
b_id integer NOT NULL,
c_id integer NOT NULL,
CONSTRAINT id PRIMARY KEY (id),
CONSTRAINT b_id FOREIGN KEY (b_id) REFERENCES "B" (id)
ON UPDATE NO ACTION ON DELETE NO ACTION, --with no action restriction
CONSTRAINT c_id FOREIGN KEY (c_id) REFERENCES "C" (id)
ON UPDATE CASCADE ON DELETE CASCADE --with cascade restriction
)
WITH (
OIDS = FALSE
)
;
ALTER TABLE "C" OWNER TO postgres;
I want to change the datatype of a primary key column in Table A, which is referenced by Table B. The schema is something like-
Table A: (col1A number, col2A...)
Table B: (col1B number, col2B...)
col2B -> col1A
I want to change datatype of col1A from number to varchar. I want that to reflect in Table B also. Is there any simple way to do that?
--
Thanks.
No, there is no simple way to do this. Assuming that both tables have data in them, you'd need to
Add a new VARCHAR2 column to table A
Update A to set the new column equal to TO_CHAR( col1A )
Add a new 'VARCHAR2` column to table B
Update B to set the column equal to TO_CHAR( col2B )
Drop the existing foreign key constraint
Drop the existing primary key constraint
Drop col1A from A
Drop col2B from B
Rename the new columns (if desired) in A & B to col1A and col2B
Create the new primary key constraint
Create the new foreign key constraint
Obviously, that's going to be a rather expensive operation.
This is a proof of concept in T-SQL (SQL Server). Basically, we're changing the primary key's data type from INT to UNIQUEIDENTIFIER (GUID) in table A, which is referenced by a foreign key in table B.
-- ARRANGE
CREATE TABLE A(
[Id] [int] IDENTITY(1,1) NOT NULL,
CONSTRAINT PK_A PRIMARY KEY (Id)
)
CREATE TABLE B(
[Id] INT IDENTITY(1,1) NOT NULL,
[A_Id] INT NOT NULL,
CONSTRAINT PK_B PRIMARY KEY (Id),
CONSTRAINT FK_B_A FOREIGN KEY (A_Id) REFERENCES A(Id)
)
INSERT A DEFAULT VALUES
DECLARE #A_Id INT
SELECT #A_Id = SCOPE_IDENTITY()
INSERT INTO B VALUES (#A_Id)
INSERT INTO B VALUES (#A_Id)
INSERT INTO B VALUES (#A_Id)
INSERT INTO B VALUES (#A_Id)
INSERT INTO B VALUES (#A_Id)
-- ACT
ALTER TABLE A ADD New_Id UNIQUEIDENTIFIER NOT NULL DEFAULT NEWID()
ALTER TABLE B ADD New_A_Id UNIQUEIDENTIFIER NULL
UPDATE B SET B.New_A_Id = (SELECT A.New_Id FROM A WHERE A.Id = B.A_Id)
ALTER TABLE B DROP FK_B_A
ALTER TABLE B DROP COLUMN A_Id
EXEC sp_RENAME 'B.New_A_Id', 'A_Id', 'COLUMN'
ALTER TABLE A DROP PK_A
ALTER TABLE A DROP COLUMN Id
ALTER TABLE A ADD CONSTRAINT PK_A PRIMARY KEY (New_Id)
EXEC sp_RENAME 'A.New_Id', 'Id', 'COLUMN'
ALTER TABLE B ADD CONSTRAINT FK_B_A FOREIGN KEY (A_Id) REFERENCES A(Id)
-- ASSERT
SELECT * FROM A
SELECt * FROM B
Can anyone please tell me how can one estabilish 1 to 0..1 and 1 to 1..* relationships between tables in SQL (Server)?
Thank you very much.
1 to 1..*
Create a Foreign key from a parent table to the primary key of the child (lookup table).
CREATE TABLE A
(
id int NOT NULL IDENTITY(1,1) PRIMARY KEY,
Somecolumn int,
SomeOtherColumn Varchar(50),
B_id int CONSTRAINT FOREIGN KEY REFERENCES B(id),
-- ...other columns
)
CREATE TABLE B
(
id int NOT NULL IDENTITY(1,1) PRIMARY KEY,
Name Varchar(50)
)
1 to 0..1
Create a table with the primary key also defined as a Foreign key to the parent table
CREATE TABLE [Master]
(
id int NOT NULL IDENTITY(1,1) PRIMARY KEY,
Somecolumn int,
SomeOtherColumn Varchar(50),
-- ...other columns
)
CREATE TABLE [Child]
(
id int NOT NULL PRIMARY KEY,
OtherColumn Varchar(50),
)
ALTER TABLE Child
ADD CONSTRAINT FK_Master FOREIGN KEY (id) REFERENCES Master(id)
One to many
Define two tables (example A and B), with their own primary key
Define a column in Table A as having a Foreign key relationship based on the primary key of Table B
This means that Table A can have one or more records relating to a single record in Table B.
If you already have the tables in place, use the ALTER TABLE statement to create the foreign key constraint:
ALTER TABLE A ADD CONSTRAINT FOREIGN KEY fk_b ( b_id ) references b(id)
* fk_b: Name of the foreign key constraint, must be unique to the database
* b_id: Name of column in Table A you are creating the foreign key relationship on
* b: Name of table, in this case b
* id: Name of column in Table B