Related tables inserting data

Question

I have a little question.

Short description of the project: I'm working on a little API project written in C#. Inside this project I'm using the onion architecture, in short it's something like Controller -> Service -> Repository.

I have an Account controller, which manages create & update User and beside this I have a databas ewith the following tables:

User, Password, UserInformation

In User I have the following columns: Id(PK), userName and role.

In Password I have: User_Id(PK and FK to User - Id column), password and oldPassword (if necessary).

In UserInformation I have: User_Id(PK and FK to User - Id column), firstName, lastName and email.

As for inserting data, I have a separate class with string constants inside which are queries and I'm executing them by doing

sqlConnection.QueryAsync(QueryFromStringConstants, etcParams).

As I'm a little bit noobish with the SQL syntax and everything related to SQL in general(haha), my question is: if I create a User , how can I insert data into all 3 tables at the same time?

Thanks:)

Answer 1

Each insert requires a separate insert statement, and you should also get an identity value after each insert in order to use it for next insert, for example:

1. insert user: INSERT INTO User (userName, role) VALUES (...)
2. get user Id (I'm sure you definded it as IDENTITY): DECLARE @userId AS integer = (SELECT @@IDENTITY);
3. insert user password (with user id obtained on prevoius step)
4. insert user info (the same)

But you can also use Entity framework. You should create class User , which have virtual properties of type Password and UserInformation (also classes you should create as well). Then create a context which has a property DbSet<User> Users . In this case you just create a User instance, populate its fields, Password and UserInformation with correspondent instances of Password and UserInformation classes. Then use:

context.Users.Add(userInstanceVar);
context.SaveChanges();

That's all, EF will save all instances into your database. and also populates all identity values automatically.

Answer 2

If your Primary Key on your User table is an int or bigint , you use the function SCOPE_IDENTITY to get the PK.

set nocount, xact_abort on;   --recommended

begin transaction;

insert into user(User_Id,userName,...)
values(...);

declare @user_id int = scope_identity();   -- make sure this is the very next line after insert

insert into Password(User_Id,....)
values(@user_id,...);

insert into UserInformation(User_Id,....)
values(@user_id,...);

commit transaction;

If you are doing a multi-row insert, you can use Table-Valued Parameters. And then you will need the OUTPUT clause to output the IDs into a new table variable:

set nocount, xact_abort on;   --recommended

begin transaction;

declare @new_users table (user_id int, username varchar(50));

insert into user(User_Id,userName,...)
output inserted.user_id, inserted.username
select ...
from @users;

insert into Password(User_Id,....)
select u.user_id, ...
from passwords p
join @new_users u on u.username = p.username;

commit transaction;

---

I must say though, that I don't see the point in having a separate table for UserInformation , after all, you cannot have a User without UserInformation , so you shouold combine these into one.

Answer 3

You have to insert in the primary table then into the others. and use transaction to run the three statements. the following is an example using oracle procedure:

create or replace PROCEDURE InsertUser(.. params ...) 
as 
temp_id number;
BEGIN
insert into user(User_Id,userName,...) values(...) returning User_Id into temp_id;
insert into Password(User_Id,....) values(temp_id,...);
insert into UserInformation(User_Id,....) values(temp_id,...);
commit;
exception when others then rollback;
END InsertUser;