Common mistakes in solving SELECT statement exercises

Tips and solutions

On the example of this simple task, a great variety of solution approaches, owing to the flexibility of SQL, can be demonstrated.

Solution 4.4.4. Outer self-join  

Console

Execute

SELECT DISTINCT p.maker 
FROM Product p LEFT JOIN 
 Product p1 ON p.maker = p1.maker AND 
 p1.type = 'Laptop'
WHERE p.type = 'PC' AND 
  p1.maker IS NULL;

SELECT DISTINCT p.maker FROM Product p LEFT JOIN Product p1 ON p.maker = p1.maker AND p1.type = 'Laptop' WHERE p.type = 'PC' AND   p1.maker IS NULL;

    Scheme

The Product table is joined to itself using LEFT JOIN on the condition the maker is the same, and the product type for the second table is  laptop. The p1.maker column will be NULL if a manufacturer doesn’t have any laptop models; this is used in the WHERE clause together with the condition the product type for the record checked is PC.

Solution 4.4.5. Grouping

Console

Execute

SELECT maker
FROM (SELECT DISTINCT maker, type 
 FROM Product 
 WHERE type IN ('PC', 'Laptop')
 ) AS a
GROUP BY maker
HAVING COUNT(*) = 1 AND 
  MAX(type) = 'PC';

SELECT maker FROM (SELECT DISTINCT maker, type FROM Product WHERE type IN ('PC', 'Laptop') ) AS a GROUP BY maker HAVING COUNT(*) = 1 AND   MAX(type) = 'PC';

    Scheme

In the subquery, unique pairs (supplier, type) are selected with type being either PC or Laptop. Then, grouping by supplier is performed, whereby the rows grouped have to meet the following conditions:

• COUNT(*) = 1 — which means, the maker has to manufacture only one product type of the selected (since printers have already been excluded by the IN predicate, it’s either PC or Laptop);
  
• MAX(type) = 'PC' — this single manufactured product type is PC. Since the HAVING clause can’t contain columns not referred to by aggregate functions, it uses MAX (type); MIN(type) would work equally well.

With so many solution approaches, it’s natural to ask about their efficiency – that is, which query will be executed faster. Leading the field both in number of operations and estimated execution time is solution 4.4.5. The third solution has the weakest performance. Others are estimated to take about twice as much time to execute as the leader.

If  A database management system (DBMS) by Microsoft Corporation. SQL(Structured Query Language) is a database computer language designed for the retrieval and management of data in relational database management systems (RDBMS), database schema creation and modification, and database object access control management.SQL Server isn’t installed on your computer, you can get the query execution plan directly from the site.

Return to discussion of exercise #8