列 'seller.seller_Name' 在 select 列表中无效，因为它不包含在聚合 function 或 GROUP BY 子句中

Question

I was creating an SQL query but ran into some problems with aggregate function or the group by clause. I've never use stack overflow before so please forgive my poor formatting and grammar. These are the codes below.

**/ Create Table Seller /

create table seller60 (
    seller_ID int identity (1000,1) not null,
    seller_Name varchar(100) not null,
    seller_userName varchar (100) unique,
    seller_password varchar (100) not null,
    item_ID int not null,
    contactID int not null,
    creditCardID int not null,
    primary key(seller_ID), foreign key(contactID) references contactInfo(contact_ID),foreign key(creditCardID) references creditCard60(creditCard_ID)
);

**/ Create Table Item /

create table item60(
    itemID int identity(1000,1) not null,
    itemName varchar(100) not null,
    itemDesc varchar (100),
    item_initialPrice money,
    item_Quantity int,
    ownerID int not null,
    condition varchar(100) not null,
    primary key( itemID), foreign key(ownerID) references seller60(seller_ID)
);

**/ The most active seller /

SELECT a.ownerID, b.seller_Name 
FROM item AS a 
INNER JOIN seller AS b ON a.ownerID = b.seller_ID 
GROUP BY a.ownerID 
ORDER BY COUNT(a.itemID) DESC set rowcount 1;

Answer 1

The problem here is that the select clause is not consistent with the group by clause: you have non-aggrgated column seller60(seller_name) in the select clause, but not in the group by clause, and MySQL rightfully complains about that.

One method is to add that column to the group by clause. A somehow smarter alternative is group by column seller60(seller_id) : this is identical to item(owner_id) (that is a foreign key to that column); and, since that is the primary key of seller60 , you can then add as many columns from that table as you want in the select clause without breaking the query (that's the concept of functionaly dependent column, which MySQL understands).

select s.* 
from item60 i
inner join seller60 s on i.ownerid = s.seller_id 
group by s.seller_id 
order by count(*) desc limit 1;

Note that this uses meaningful table aliases, rather than random letters such as a or b . I also fixed the limit clause, whose syntax was off.

Finally: you could also get the same result with a correlated subquery - and then you don't need to worry about aggregation at all.

select s.*
from seller60 s
order by (select count(*) from item60 i where i.ownerid = s.seller_id) desc limit 1