LeetCode 584 - Find Customer Referee¶
Database Language: SQL Server
Difficulty:
Problem Description¶
Input¶
Table: Customer¶
Column Name | Type |
---|---|
id | int |
name | varchar |
referee_id | int |
In SQL, id
is the primary key column for this table.
Each row of this table indicates the id of a customer, their name, and the id of the customer who referred them.
Requirement¶
Find the names of the customer that are not referred by the customer with id = 2.
Return the result table in any order.
The result format is in the following example.
Examples¶
Example 1¶
Input¶
Customer table:
id | name | referee_id |
---|---|---|
1 | Will | NULL |
2 | Jane | NULL |
3 | Alex | 2 |
4 | Bill | NULL |
5 | Zack | 1 |
6 | Mark | 2 |
Output¶
name |
---|
Will |
Jane |
Bill |
Zack |
SQL Schema¶
CREATE TABLE Customer (id INT PRIMARY KEY, name VARCHAR(25), referee_id INT);
TRUNCATE TABLE Customer;
INSERT INTO Customer (id, name, referee_id) values ('1', 'Will', NULL);
INSERT INTO Customer (id, name, referee_id) values ('2', 'Jane', NULL);
INSERT INTO Customer (id, name, referee_id) values ('3', 'Alex', '2');
INSERT INTO Customer (id, name, referee_id) values ('4', 'Bill', NULL);
INSERT INTO Customer (id, name, referee_id) values ('5', 'Zack', '1');
INSERT INTO Customer (id, name, referee_id) values ('6', 'Mark', '2');
Solutions¶
There are three possible ways of finding the names of customers that are not referred by the customer with id = 2, namely:
- Using the
OR
operator - Using the
ISNULL
system function - Using the
COALESCE
comparison function
Solution 1 - Using OR Operator¶
To find the names of customers that are not referred by the customer with id = 2, all that needs to be done is check if the referee_id
is not equal to 2. This can be achieved by the following query:
SELECT name FROM Customer
WHERE referee_id != 2
Running this query returns the following result, which is not the required result:
name |
---|
Zack |
The query is not returning those customers where the referee_id
is NULL because NULL != 2
returns an unknown value. According to SQL Server - NULL and UNKNOWN (Transact-SQL),
NULL indicates that the value is unknown. A null value is different from an empty or zero value. No two null values are equal. Comparisons between two null values, or between a null value and any other value, return unknown because the value of each NULL is unknown.
To include customers that were not referred by any other customers, referee_id IS NULL
needs to be added in the query. The condition referee_id = NULL
cannot be used because as mentioned above, the result of any arithmetic comparison, such as =
, with NULL is unknown. Adding the referee_id IS NULL
condition to the query above now yields the following query:
# Final Solution Query
SELECT name FROM Customer
WHERE referee_id != 2 OR referee_id IS NULL
This query now returns the desired result. Here's the query plan generated by SQL Server for this query:
|--Clustered Index Scan(OBJECT:([leetcode].[dbo].[Customer].[PK_Customer]),
WHERE:([leetcode].[dbo].[Customer].[referee_id] IS NULL OR [leetcode].[dbo].[Customer].[referee_id]<>(2)))
And here's the fastest runtime for this query:
- Runtime: 462ms
- Beats: 93.82% as of July 27, 2024
One may ask, does the order of the conditions used by the OR
operator matter? So instead of WHERE referee_id != 2 OR referee_id IS NULL
, will it be faster if the conditions were interchanged to WHERE referee_id IS NULL OR referee_id != 2
? Will the following query be faster than the one earlier?
# Final Solution Query
SELECT name FROM Customer
WHERE referee_id IS NULL OR referee_id != 2
Comparing the query plan of this updated query with the one above, it can be seen that the same query plan was generated by SQL Server:
|--Clustered Index Scan(OBJECT:([leetcode].[dbo].[Customer].[PK_Customer]),
WHERE:([leetcode].[dbo].[Customer].[referee_id] IS NULL OR [leetcode].[dbo].[Customer].[referee_id]<>(2)))
In addition, the fastest runtime for the "interchanged" condition is somewhat close to the one above:
- Runtime: 462ms
- Beats: 93.82% as of July 27, 2024
There have been a lot of discussions in the SQL Server community about whether SQL Server performs a "short-circuit" when evaluating the WHERE
clause that involves the OR
operator and no "official" documentation can be found that neither confirms nor denies that SQL Server performs a "short-circuitwhen evaluating the
WHEREclause that involves the
OR` operator.
One other question that may be asked, given that the referee_id
is being compared between 2 values, namely the value 2 and NULL, can't the NOT IN
comparison operator be used:
SELECT name FROM Customer
WHERE referee_id NOT IN (NULL, 2);
Running this query will return an empty set. The reason the query is returning an empty set is because according to SQL Server - IN Logical Operator (Transact-SQL):
Any null values returned by subquery or expression that are compared to test_expression using IN or NOT IN return UNKNOWN. Using null values in together with IN or NOT IN can produce unexpected results.
Since one of the expressions or values in the list is a NULL value, the result of the condition returns an UNKNOWN value. Thus, the referee_id NOT IN (NULL, 2)
condition returns False for each row in the table because in SQL Server, 0, NULL or UNKNOWN means false and anything else means true.
Solution 2 - Using ISNULL System Function¶
The second way of finding the names of customers who were not referred by the customer with id = 2 is with the use of the ISNULL
system function. The ISNULL
function accepts 2 parameters and returns the first parameter if it is not NULL; otherwise it returns the second parameter.
Here's how the query will now look using the ISNULL
flow control function instead of the OR
operator:
# Final Solution Query
SELECT name FROM Customer
WHERE ISNULL(referee_id, 0) != 2
With the ISNULL(referee_id, 0)
, if the value of referee_id
is not NULL, then the function returns the value of referee_id
. Otherwise, if the value of referee_id
is NULL, the function returns the second parameter, which is in this case, 0. The value passed as the second parameter of the ISNULL
function, in this case, can be any number as long as it is not 2 so that if referee_id
is NULL, the WHERE
clause will return a true value.
Here's the query plan generated by SQL Server:
|--Clustered Index Scan(OBJECT:([leetcode].[dbo].[Customer].[PK_Customer]),
WHERE:(isnull([leetcode].[dbo].[Customer].[referee_id],(0))<>(2)))
And here's the fastest runtime for this query using the ISNULL
function:
- Runtime: 462ms
- Beats: 93.82% as of July 27, 2024
Solution 3 - Using COALESCE Comparison Function¶
The third way of finding the names of customers who were not referred by the customer with id = 2 is with the use of the COALESCE
comparison function. The COALESCE
function returns the first non-NULL value in the list or NULL if there are no non-NULL values in the list.
Here's how the query will look like using the COALESCE
comparison function instead of the ISNULL
flow control function or the OR
operator:
# Final Solution Query
SELECT name FROM Customer
WHERE COALESCE(referee_id, 0) != 2
Similar to the ISNULL
function in the second solution, with the COALESCE(referee_id, 0)
, if the value of referee_id
is not NULL, then the COALESCE
function returns the value of referee_id
because that's the first non-NULL value in the list of parameters passed to the COALESCE
function. Otherwise, if the value of referee_id
is NULL, the COALESCE(referee_id, 0)
function returns 0 because that's the first non-NULL value in the list of parameters passed to the function. Just like the case for the ISNULL
function in the second solution, the second parameter passed to the COALESCE
function can be any number aside from 0 as long as it is not 2 so that the WHERE
clause will return a true value if referee_id
is NULL.
Here's the query plan generated by SQL Server:
|--Clustered Index Scan(OBJECT:([leetcode].[dbo].[Customer].[PK_Customer]),
WHERE:(CASE WHEN [leetcode].[dbo].[Customer].[referee_id] IS NOT NULL
THEN [leetcode].[dbo].[Customer].[referee_id] ELSE (0) END<>(2)))
And here's the fastest runtime for the query that uses the COALESCE
function:
- Runtime: 464ms
- Beats: 92.61% as of July 27, 2024
Solution Runtime Comparison¶
Here's the comparison of the fastest runtime for each of the solutions.
Solution # | Runtime | Beats |
---|---|---|
1a - Using OR Operator - NULL Check Last | 462ms | 93.82% |
1b - Using OR Operator - NULL Check First | 462ms | 93.82% |
2 - Using ISNULL Function | 462ms | 93.82% |
3 - Using COALESCE Function | 464ms | 92.61% |