Question

Following is a table structure in my system. The scenario is, system has Clients. Clients can be
many type, for exampel: a Person, a Company, a Bank or etc. I have come with a database design as follows.

TABLE : CLIENT

===============================
| ID    | NAME_CODE | TYPE    |
-------------------------------
| 1000  | Martin    | PERSON  |
| 1002  | George    | PERSON  |
| 1003  | Max_Group | COMPANY |
-------------------------------

TABLE : PERSON

==================================================
|  ID | CLIENT_ID  | FIRST_NAME   | LAST_NAME    |
--------------------------------------------------
| 1   | 1001       | Martin       | XXXXXXXXXXX  |
| 2   | 1002       | George       | XXXXXXXXXXX  |
--------------------------------------------------

TABLE : COMPANY

===================================================
| ID | CLIENT_ID  | NAME           | Location     |
---------------------------------------------------
| 1  | 1003       | Max Group Inc. | XXXXXXXXXXX  |
---------------------------------------------------

EDIT : ID added for PERSON and COMPANY tables.

I welcome if any one could suggest me a better structure.

For the above structure, I need to creat the class model. Following is what I have created.

class Client
{
   int ID;
   string NameCode;
}

class Person : Client
{
    int ID;
    int ClientId;
    string FirstName;
    stirng LastName;
    //........
}

class Company : Client
{
    int ID;
    int ClientId;
    string Name;
    string Location;
    //.........
}

I feel the above model are not suitable. There can be a right way for this.
Please suggest me a better way.

Answer 1

After some discussion, I think your tables are well designed.

I would design your tables this way:

TABLE : CLIENT

==============================
| ID   | NAME_CODE | TYPE    |
------------------------------
| 1    | Martin    | 1       |
| 2    | George    | 1       |
| 3    | Max_Group | 2       |
------------------------------

where Type is an Enum rather than anything hard-coded so that queries are faster.

TABLE : PERSON

============================================
| CLIENT_ID  | FIRST_NAME   | LAST_NAME    |
--------------------------------------------
| 1          | Martin       | XXXXXXXXXXX  |
| 2          | George       | XXXXXXXXXXX  |
--------------------------------------------

TABLE : COMPANY

==============================================
| CLIENT_ID  | NAME           | Location     |
----------------------------------------------
| 3          | Max Group Inc. | XXXXXXXXXXX  |
----------------------------------------------

And the classes would look like:

class Client
{
   int ID;
   string NameCode;
    //........
}

class Person : Client
{
    string FirstName;
    stirng LastName;
    //........
}

class Company : Client
{
    string Name;
    string Location;
    //.........
}

The Client table has the parent id now for every client entity. It could have been reversed by having individual id for persons and banks and then reference a foreign key in the parent client table which I believe is easier to operate from db point of view (that scales well if you’re going for composition). But in the application I love inheritance and polymorphism, so the straight adaptation of it for db design is what I posted.