为什么count（column_name）的行为如此？

Question

Consider the relation account (customer, balance) where customer is a primary key and there are no null values.

I would like to rank customers according to decreasing balance. (The customer with the largest balance gets rank 1. Ties are not broke but ranks are skipped: if exactly two customers have the largest balance they each get rank 1 and rank 2 is not assigned.)

Why does the following query never prints the customer having rank 1?

    select A.customer, 1+count(B.customer) 
    from account A, account B 
    where A.balance < B.balance 
    group by A.customer 

SQLfiddle Link

Answer 1

Let's try this Example

Ranking without skipping a rank

set @number:=0;
set @balance:=0;

select customer, balance, rank 
from (
  select 
    *, 
    @number:=if(@balance=balance, @number, @number+1) as rank,
    @balance:=balance
  from account
  order by balance
) as rankeddata;

Result

customer balance rank
S        300     1
Q        400     2
R        400     2
P        500     3

To show ranking from 500 -> 300, just change the ORDER BY balance to ORDER BY balance DESC

Skip rank if multiple rows have same rank

If you prefer to skip an assigned rank, you can tweak the query a little bit SQL Fiddle .

set @number:=0;
set @balance:=0;
set @rank_reused:=0;

select customer, balance, rank
from (
  select 
    *, 
    @number:=if(@balance=balance, @number, @number+1+@rank_reused) as rank,
    @rank_reused:=if(@balance=balance, @rank_reused+1, 0) as reused,
    @balance:=balance
  from account
  order by balance desc
) as rankeddata;

Result

customer balance rank
S        300     1
Q        400     2
R        400     2
P        500     4

Answer 2

This has nothing to do with the count.

You are doing a join between A and B and show only the records where A.balance < B.balance since there is no such record for your top rank customer (by definition there is no account with higher balance) you don't get any record.

This should do the trick

select A.customer, ( 
    select count(*) + 1 
    from account B
    where A.balance < B.balance
) from account A

Answer 3

The following statement will give you a rank of all customers based on their balance:

SELECT A.customer, A.balance, ( 
    SELECT count(*)+1 
    FROM account B
    WHERE A.balance < B.balance
  ) AS rank
FROM account A ORDER BY A.balance DESC

The result will be:

customer balance rank
P   500 1 
Q   400 2 
R   400 2 
S   300 4

order by A.balance does the ordering, while desc ranks from largest to smallest. Lastly, the subquery provides the rank (note the +1).

Here's an the example fiddle .

Answer 4

Try

select A.customer, 
(select count(*) from account) -
(select count(*) from account a1 where a1.balance<a.balance) rank
from account A
Order by rank