来自SQL中多个列的DISTINCT值

Question

I have the following assignment:

Write a SELECT statement that returns one row for each customer that has the same last name and billing address as another customer. Sort the result set by last name then first name.

I have tried doing it with the DISTINCT keyword but that does not serve my purpose.

For some reason every time I use GROUP BY I get the following error:

SELECT FirstName, LastName, BillingAddressID
    FROM Customers
    GROUP BY LastName;

Column 'Customers.FirstName' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

I tried UNION with the following error:

SELECT LastName FROM Customers

UNION 

SELECT BillingAddressID FROM Customers

Conversion failed when converting the varchar value 'Sherwood' to data type int.

Can someone just point me in the right direction?

This is a sample of the data set that I am working on

firstname   lastname    billingaddressid  
Allan       Sherwood        2  
Barry       Zimmer          3  
Christine   Brown           4  
David       Goldstein       6  
Erin        Sherwood        7  
Frank Lee   Wilson          8  
Gary        Brown           4  
Heather     Esway           12  

So, the query should remove the duplicate entries..of the lastname

Answer 1

Try -

SELECT FirstName,
       Customers.LastName,
       Customers.billingAddressID
FROM Customers INNER JOIN
     ( SELECT LastName,
              billingAddressID
       FROM Customers
       GROUP BY LastName,
                billingAddressID
       HAVING COUNT( LastName ) >= 2 ) lastNameQuery
  ON Customers.LastName = lastNameQuery.LastName AND
     Customers.billingAddressID = lastNameQuery.billingAddressID
ORDER BY Customers.LastName,
         FirstName;

I tested this against a database created using the following script...

CREATE DATABASE Cust;

USE Cust;

CREATE TABLE Customers
(
    fldID               INT              NOT NULL    AUTO_INCREMENT,
    firstName           VARCHAR( 50 )    NOT NULL,
    lastName            VARCHAR( 50 ),
    billingAddressID    INT              NOT NULL,
    PRIMARY KEY ( fldID )
);

I entered the Questioner's sample data using -

INSERT INTO Customers
SET firstName        = "Allan",
    lastName         = "Sherwood",
    billingAddressID = 2;

INSERT INTO Customers
SET firstName        = "Barry",
    lastName         = "Zimmer",
    billingAddressID = 3;

INSERT INTO Customers
SET firstName        = "Christine",
    lastName         = "Brown",
    billingAddressID = 4;

INSERT INTO Customers
SET firstName        = "David",
    lastName         = "Goldstein",
    billingAddressID = 6;

INSERT INTO Customers
SET firstName        = "Erin",
    lastName         = "Sherwood",
    billingAddressID = 7;

INSERT INTO Customers
SET firstName        = "Frank Lee",
    lastName         = "Wilson",
    billingAddressID = 8;

INSERT INTO Customers
SET firstName        = "Gary",
    lastName         = "Brown",
    billingAddressID = 10;

INSERT INTO Customers
SET firstName        = "Heather",
    lastName         = "Esway",
    billingAddressID = 12;

I also added the following to ensure a repeat of BOTH lastName AND billingAddressID...

INSERT INTO Customers
SET firstName        = "Don",
    lastName         = "Sherwood",
    billingAddressID = 22;

INSERT INTO Customers
SET firstName        = "Timmy",
    lastName         = "Sherwood",
    billingAddressID = 22;

INSERT INTO Customers
SET firstName        = "James",
    lastName         = "Brown",
    billingAddressID = 22;

INSERT INTO Customers
SET firstName        = "James",
    lastName         = "Esway",
    billingAddressID = 22;

The question being asked of our Questioner, and I assume the question that the Questioner is seeking help with, was -

Write a SELECT statement that returns one row for each customer that has the same last name and billing address as another customer. Sort the result set by last name then first name.

My interpretation of this was we should return records for EACH Customer meeting the criteria of having a particular combination BOTH a Last Name AND a BillingAddressID shared with AT LEAST one other Customer, and that the list of returned records should be sorted by Last Name and subsorted on First Name.

The core of my answer is the segment -

SELECT LastName,
       billingAddressID
FROM Customers

Which selects just the two conditional fields from Customers.

To this I added -

GROUP BY LastName,
         billingAddressID

This will refine the core segments results into a list of each unique combination of the two conditional fields.

I then restricted this list to those unique combinations that occur at least twice by adding -

HAVING COUNT( LastName ) >= 2

I then gave the resulting query an Alias of lastNameQuery .

I then joined lastNameQuery with Customers on both conditional fields to restrict our list of Customers to those who had a shared pair of conditional values using -

     Customers INNER JOIN
     ( SELECT LastName,
              billingAddressID
       FROM Customers
       GROUP BY LastName,
                billingAddressID
       HAVING COUNT( LastName ) >= 2 ) lastNameQuery
  ON Customers.LastName = lastNameQuery.LastName AND
     Customers.billingAddressID = lastNameQuery.billingAddressID

From this I selected the desired fields using -

SELECT FirstName,
       Customers.LastName,
       Customers.billingAddressID
FROM

Using Customers. was necessitated by the need to clarify references to field names that occur in both Customers and lastNameQuery . Without this clarification MySQL is unable to determine which fields it should use.

This list was sorted into the specified order by adding -

ORDER BY Customers.LastName,
         FirstName;

The results I got from testing my complete statement were -

+-----------+----------+------------------+
| FirstName | LastName | billingAddressID |
+-----------+----------+------------------+
| Gary      | Brown    |               10 |
| Tom       | Brown    |               10 |
| Don       | Sherwood |               22 |
| Timmy     | Sherwood |               22 |
+-----------+----------+------------------+

These are the only Customers in my expanded sample dataset to have a shared combination of values in the conditional fields.

Answer 2

Try this

SQL Fiddle

   SELECT * FROM 
    (
    SELECT *,ROW_NUMBER() OVER(PARTITION BY LASTNAME ORDER BY FIRSTNAME) AS RN
    FROM YOURTABLE
    )T
    WHERE RN=2

If you want to return all unique records.

SELECT * FROM 
    (
    SELECT *,ROW_NUMBER() OVER(PARTITION BY LASTNAME ORDER BY FIRSTNAME) AS RN
    FROM YOURTABLE
    )T
    WHERE RN=1