Question

I have a database that I made for a website I am making that will have a list of employees. It will be a grid like list with each employee pic, name, phone number, etc.

This is my first database design and I’m really proud of it… but I’m a newbie so I am sure it is inefficient in some ways.

I am having trouble figuring out the best way to query it for the website to make the elements.

1) Should I use a virtual table? Or should I just use one query and loop it for each employee?

2) Some employees will have more than one jurisdiction. Right now I have:

select employees.firstName, employees.lastName, employees.phone, employees.position,
    employees.picture, jurisdictions.jurisdiction
from employees
    inner join(jurisdictions cross join departments cross join user_jurisdictions)
    on (employees.deptID = departments.deptID
        AND user_jurisdictions.userID = employees.userID
        AND user_jurisdictions.jurID = jurisdictions.jurID)

Which would return:

+-----------+----------+--------------+-----------------+---------+--------------+
| firstName | lastName | phone        | position        | picture | jurisdiction |
+-----------+----------+--------------+-----------------+---------+--------------+
| John      | Smith  | 210-226-3232 | Senior Manager   |/pics/jpeg1|South America|
| John      | Smith  | 210-226-3232 | Senior Manager   |/pics/jpeg1|London       |
+-----------+----------+--------------+-----------------+---------+--------------+

How is there a way to create the query where it would remove the duplicates rows because of the additional jurisdiction the employee would have? something like?

+-----------+----------+--------------+---------+---------+--------------+------------+
| firstName | lastName | phone        | position| picture | jurisdiction |Jurisdiction|
+-----------+----------+--------------+---------+---------+--------------+------------+
| John      | Smith  | 210-226-3232 | Manager |/pics/jpeg1| South America|London      |
+-----------+----------+--------------+-----------------+---------+-------------------+

Here is my mysql schema:

employees
--------------------
userID (primary key)
firstName
lastName
phone
fax
position
picture
deptID (foreign key references departments(deptID)) 



departments
--------------------
deptID (primary key)
department

jurisdictions
-----------------
jurID (primary key)
jurisdiction


user_jurisdictions
--------------------------
userID(foreign key references employees(userID)) 
jurID(foreign key references jurisdictions(jurID))


triger: 
 after_employees_insert | INSERT | employees | begin
insert into user_jurisdictions(userID) values (new.userID);

Answer 1

You can group your results by employee and use MySQL’s GROUP_CONCAT() function to aggregate the employee’s jurisdictions:

SELECT   employees.firstName,
         employees.lastName,
         employees.phone,
         employees.position,
         employees.picture,
         GROUP_CONCAT(jurisdictions.jurisdiction)
FROM     employees
    JOIN departments        USING (deptID)
    JOIN user_jurisdictions USING (userID)
    JOIN jurisdictions      USING ( jurID)
GROUP BY employees.userID

However, this aggregates jurisdictions into a delimited string (which makes it difficult to differentiate between multiple jurisdictions and a single jurisdiction containing the delimiter: one can choose a delimiter that is unlikely to be used in a jurisdiction name, but it’s still subject to problems such as limitations in string length and, in my view, is fundamentally lazy).

A better way might be to leave such aggregation to a higher-level of your application code. For example, using PDO:

$dbh = new PDO("mysql:dbname=$dbname", $username, $password);

$qry = $dbh->query('
  SELECT   employees.userID
           employees.firstName,
           employees.lastName,
           employees.phone,
           employees.position,
           employees.picture,
           jurisdictions.jurisdiction
  FROM     employees
      JOIN departments        USING (deptID)
      JOIN user_jurisdictions USING (userID)
      JOIN jurisdictions      USING ( jurID)
  ORDER BY employees.userID
');

$row = $qry->fetch();
while ($row) {
  $current_user = $row['userID'];

  // handle $current_user initialisation
  do {
    // handle user $row
  } while ($row = $qry->fetch() and $row['userID'] == $current_user);
  // handle $current_user termination

}

Of course, this has the disadvantage of producing a much larger resultset that must be generated by the RDBMS and transferred to / processed by your higher-level code.