SQL Subquery or Group BY to return Distinct values of a certain column? - sql

Question

I am running into an issue where I know I am close to a solution but I cannot get the right data to come back or am getting errors on the queries.
Basically what I want to do is to select values from a single table based on the following: I want them based on an employee name I supply and I want them to return DISTINCT Account Numbers for that employee(sometimes the same account number is listed more than one time...I only want one row of data from each duplicate account number).
Here is what I have tried so far:
SELECT DISTINCT * FROM Employees WHERE EmpName = 'Mary Johnson' GROUP BY AccountID
-This returns an error saying I don't have an aggregate for the column EmployeeID(which isn't a calculated column, so why is it asking for an aggregate??)
Then I tried a subquery:
SELECT EmployeeID, EmpName, Department, Position, (SELECT DISTINCT AccountID From Employees) AS AcctID FROM Employees WHERE EmpName = 'Mary Johnson'
-This gives me an error saying AT Most one record can be returned(I'm in Access).
I know there has to be a solution to what I am looking for---Currently If I do a full query just on EmpName it returns 30 records, however, 23 of them have a duplicate AccountID, so there should only be 7 records that return with a DISTINCT AccountID.
How can I achieve this via SQL?
Here is an Example:
ID(PK) EmpName AcctID Department Position EmpId
------ -------- --------- -------- -------- ------
1 Mary Johnson 1234 IT Tech 226663
2 Mary Johnson 1234 IT Tech 226663
3 Mary Johnson 1234 IT Tech 226663
4 Mary Johnson 2345 IT Tech 226663
5 John James 23442 Banking Teller 445645
6 Jame Tabor 1234 HR Manager 234555
In the example above I want to do an SQL Query to get the rows back with the AccountId's for Mary Johnson that are DISTINCT. So We would get back 2 rows Mary Johnson 1234 and Mary Johnson 2345, ignoring the other 2 rows that Mary Johnson has with duplicate AccountIDs of 1234 and the other rows with employees not named Mary Johnson

Answer 1

This is not really easy to do in MS Access -- it lacks row_number(). One method is to use another column that uniquely identifies each row that has the same AccountID:
select e.*
from Employees as e
where e.?? = (select max(e2.??)
from Employees as e2
where e2.AccountId = e.AccountId and e2.empName = e.empName
);
The ?? is for the column that uniquely identifies the rows.
If you only care about two columns, then use select distinct:
select distinct e.empName, e.AccountId
from employees as e;
You an add a where clause to either query to restrict to a single employee. However, it doesn't make sense to me that a table called Employees would have multiple rows for a single employee name (except in the rare case of people with the same names).

Answer 2

The reason you are getting the error about "you can only return one row" is because you are using a subquery to feed into a record row. In that case, you are searching for distinct AcctIDs for Mary Johnson, she has two distinct ones, so the query fails. If you used MIN(AcctID) that query would work fine. Therefore, this will work:
SELECT EmployeeID, EmpName, Department, Position, MIN(AccountID )
FROM Employees
WHERE EmpName = 'Mary Johnson'
GROUP BY EmployeeID, EmpName, Department, Position
The next question you need to ask yourself is WHICH AccountID do you want to see? You can switch to MAX, but if you have three values and you want the middle one, that will be a lot harder to solve. But, the above query will at least give you a working result.
EDIT:
That should give you only one row assuming you have identical information for Mary Johnson. If you are getting 24 it means you have different entries For empname, department, position, or employeeid. If you want to see all AccountIDs associated with Mary Johnson, use this:
SELECT EmployeeID, MIN(EmpName), AccountID, MIN(Department) AS Department, MIN(Position) AS Position
FROM Employees
WHERE EmpName = 'Mary Johnson'
GROUP BY EmployeeID, AccountID