Conditional SQL - mysql

Question

I just recently started learning how to write SQL queries, and I have a lot to learn and a lot of questions, mainly regarding defining conditions for an SQL query. I have 3 tables (with fields listed below):
Employee:
EmployeeID, Name, DoB, StartDate
Salary:
SalaryID, DataPaid, AmountPaid, EmployeeID
Address:
AddressID, Address, City, EmployeeID
Now I would like to know how to:
1. Count the number of employees that live in the city of London.
My attempt:
SELECT COUNT(City) AS EmployeeID FROM Address
WHERE City='London';
2. Add up the 'AmountPaid' values for all employees from the city of London.
My attempt:
SELECT SUM(AmountPaid) AS TotalPaid FROM Salary
WHERE City='London';
3. Display data for all employees that started in 2012 (have a 'Start Date' containing 2012). Not sure where to begin with this!
4. Delete all records where the employee 'Name' field is empty/null.
My attempt:
DELETE FROM Employee
WHERE Name=NULL;
Am I doing something wrong with my attempts? Any help would be appreciated.
THANK YOU!

Answer 1

In SQL Server (T-SQL), you cannot test a value for NULL using '='. It must be as follows:
Delete From Employee Where Name IS NULL

Answer 2

Also i would check for empty names:
DELETE FROM Employee where Name IS NULL OR Name=''
As for the point 3:
SELECT * FROM Employee WHERE Year(StartDate)=2012
As for the point 2, the table Salary does not have an City column. You need to join with Employee table like this:
SELECT SUM(AmountPaid) AS TotalPaid FROM Salary SA inner join Employee Em on Em.EmployeeID=SA.EmployeeID WHERE Em.City='London';

Answer 3

Smells a bit like homework. Anyway:
(1) count of employees in the city of London
First statement seems correct to me, isn't it? But the alias confused me a bit as it says EmployeeID but returned value is the count of employees.
(2) total amount paid for employees in the city of London
This wouldn't work with your statement because table Salary has no field City. You'll need a join here:
select sum(AmountPaid) as TotalPaid
from Salary s
join Address a on s.EmployeeID = a.EmployeeID
where a.City = 'London'
(3) employees started in 2012
Here you can use YEAR function to extract the year out of the date:
select *
from Employee
where Year(StartDate) = 2012
(4) delete unnamed employees
Taken from the solution of #ericpap (for the sake of completeness):
delete from Employee
where Name is NULL
or Name = ''

Answer 4

I would do it like this - it includes a test whereby you make tables and data to help you see how it works;
I assume a employee has only one address, but recieves multiple payments and while it may not be necessary to join from the employee tables in your example - it shows how the relationships between employees and the other tables work, which may help you understand the relational model better.
--Make the Tables
CREATE TABLE dbo.Employee
(
EmployeeID BIGINT PRIMARY KEY NOT NULL,
Name VARCHAR(50),
DoB DATE,
StartDate DATE
)
CREATE TABLE dbo.Salary
(
SalaryID BIGINT PRIMARY KEY NOT NULL,
DatePaid DATE, -- I think you mean DatePaid and not DataPaid
AmountPaid MONEY,
EmployeeID BIGINT
)
CREATE TABLE dbo.Address
(
AddressID BIGINT PRIMARY KEY NOT NULL,
[Address] VARCHAR(max),
City VARCHAR(200),
EmployeeID BIGINT
)
-- Put in some Test Data
INSERT INTO dbo.Employee (EmployeeID,Name, DoB, StartDate)
VALUES (1,'Bill Gates','19551028','20121014'),
(2,'Larry Ellison','19440817','20140101')
INSERT INTO dbo.Address (AddressID,[Address], City, EmployeeID)
VALUES (1,'15 Microsoft House','New York',1),
(2,'23 Oracle Flats','London',2)
INSERT INTO dbo.Salary(SalaryID, DatePaid, AmountPaid, EmployeeID)
VALUES (1,Getdate(),5000.53,1),
(2,'20140201',10000.23,2),
(3,'20140301',10000.23,2)
-- Queries;
--Count the number of employees that live in the city of London.
SELECT COUNT(Distinct E.EmployeeID) as Count_London
FROM dbo.Employee E
INNER JOIN dbo.[Address] A
ON E.EmployeeID = A.EmployeeID
WHERE A.City = 'London'
-- Result = 1
-- 2. Add up the 'AmountPaid' values for all employees from the city of London
--Total Amount
SELECT Sum(S.AmountPaid) as TotalPaid
FROM dbo.Employee E
INNER JOIN dbo.[Address] A
ON E.EmployeeID = A.EmployeeID
LEFT JOIN dbo.Salary S
ON E.EmployeeID = S.EmployeeID
WHERE A.City = 'London'
-- Result = 20000.46 (2 x 10000.23)
--split by employee;
SELECT E.EmployeeID,E.Name,Sum(S.AmountPaid) as TotalPaid
FROM dbo.Employee E
INNER JOIN dbo.[Address] A
ON E.EmployeeID = A.EmployeeID
LEFT JOIN dbo.Salary S
ON E.EmployeeID = S.EmployeeID
WHERE A.City = 'London'
GROUP BY E.EmployeeID, E.Name
--3. Display data for all employees that started in 2012
SELECT *
FROM dbo.Employee E
INNER JOIN dbo.[Address] A
ON E.EmployeeID = A.EmployeeID
LEFT JOIN dbo.Salary S
ON E.EmployeeID = S.EmployeeID
WHERE StartDate >= '20120101' AND StartDate < '20130101'
-- result = all bill gates fields
-- 4. Delete all records where the employee 'Name' field is empty/null.
DELETE FROM dbo.Employee WHERE EmployeeID IS NULL
you might want to delete where the employeeID is null from all the tables if it is possible to have records in there

Answer 5

This wasn't addressed in the other answers:
Count the number of employees that live in the city of London
Your attempt below will get you the right answer only if there is a different employee per row.
SELECT COUNT(City) AS EmployeeID FROM Address
WHERE City='London';
If your Address table looks like it does below, using the query you wrote, you would get 3 employees (but we can see from the table there are only 2 distinct EmployeeID's so thus only 2 employees).
AddressID, Address, City, EmployeeID
1, 1 Main St, London, 1
2, 2 Main St, London, 2
3, 3 Main St, London, 1
What you should do is count the number of distinct EmployeeID's. Each of these EmployeeID's identifies a unique employee. Thus your query should look like this:
SELECT COUNT(DISTINCT EmployeeID) AS EmployeeID FROM Address
WHERE City='London';
That will get you the correct answer every time regardless of what your table looks like.