T-SQL GROUP BY nvarchar 错误

Question

I was wondering if anyone could help me with a request kindly:

Here is the data:

Table 1: EMPLOYEE

    FK: DID
    PK: Name 
    UserName

Table 2: DEPARTMENT

    PK: DID
    TerminationDate

I'm looking to find the number of terminated employees in the quarter. Here is the T-SQL so far:

SELECT 
    DEPARTMENT.name AS Name,
    COUNT(e.userName)
FROM
    EMPLOYEE AS e
JOIN 
    DEPARTMENT ON e.department = DEPARTMENT.DID

UNION 

SELECT
    u.eu, u.name 
FROM 
    (SELECT 
         dd.name, COUNT(ee.userName) AS eu
     FROM 
         DEPARTMENT AS dd
     JOIN 
         EMPLOYEE AS ee ON dd.DID = ee.department
                        AND ee.terminationDate IS NOT NULL
     WHERE
         ee.terminationDate IS NOT NULL
         AND ee.terminationDate BETWEEN '2015-04-01' AND '2015-06-30'
     GROUP BY 
         dd.name, ee.userName) AS u
GROUP BY 
    u.eu, u.name, Name      
ORDER BY 
    Name

The error is:

Msg 8120, Level 16, State 1, Line 1
Column 'DEPARTMENT.name' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

Answer 1

In the first part of your query, you list the department name in your select clause but don't include it in a group by clause (because it is missing)

select DEPARTMENT.name as Name, COUNT(e.userName)
from EMPLOYEE as e
join DEPARTMENT on e.department = DEPARTMENT.DID
group by DEPARTMENT.Name --add this line

Answer 2

The number of problems with this query are bewildering. To start with... The table definitions you have listed do not match your query. You have the TerminationDate as an attribute of the DEPARTMENT table, where as in the query it looks like you are accessing it though the EMPLOYEE table, which makes sense to be there. So I'm going to assume that your table structure really is:

Table 1: EMPLOYEE

FK: DID
PK: Name 
UserName
TerminationDate

Table 2: DEPARTMENT

PK: DID
Name  

The way you have used a UNION in this query leads me to believe that you have a misunderstanding of what a UNION statement does. UNION is used to join two sets into one. So in set theory {1,2,3} union {3, 4} is {1,2,3,3,4}. In SQL Server, you have to make sure that the types and number of columns in the first select statement match all the selects that are unioned with it. So for example:

 Select 1 A, 2 B
 Union
 Select 3 B

Will give this error because the first select has two columns and the second has only one:

Msg 205, Level 16, State 1, Line 3 All queries combined using a UNION, INTERSECT or EXCEPT operator must have an equal number of expressions in their target lists.

And this query:

 Select 1 A
 Union
 Select 'Bob' A

Will give this error because the column types do not match:

Msg 245, Level 16, State 1, Line 2 Conversion failed when converting the varchar value 'Bob' to data type int.

If all you are looking for is the number of employees terminated between two dates, then there really isn't really anything you need to union. All the employees are already together in one set (the Employee table)... and you just need to filter the dates you want and count the record total. You don't even need the DEPARTMENT table to calculate this.

I could write the query for you... but I'll leave the rest to you.