更新基于总组的秩列

Question

我写了下面的代码SQLite的：

CREATE TABLE payments (
    customerNumber REAL NOT NULL, 
    checkNumber TEXT NOT NULL, 
    paymentDate TEXT NOT NULL, 
    amount REAL NOT NULL, 
    rank REAL default 0, 
    PRIMARY KEY (customerNumber,checkNumber) 
); 

insert into payments(customerNumber,checkNumber,paymentDate,amount) values(103,'HQ336336','2004-10-19 00:00:00',6066.78); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(103,'JM555205','2003-06-05 00:00:00',14571.44); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(112,'BO864823','2004-12-17 00:00:00',14191.12); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(112,'HQ55022','2003-06-06 00:00:00',32641.98); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(121,'MA302151','2004-11-28 00:00:00',34638.14); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(121,'KI831359','2004-11-04 00:00:00',17876.32); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(161,'KG644125','2005-02-02 00:00:00',12692.19); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(161,'NI908214','2003-08-05 00:00:00',38675.13); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(181,'CM564612','2004-04-25 00:00:00',22602.36); 
insert into payments(customerNumber,checkNumber,paymentDate,amount) values(181,'GQ132144','2003-01-30 00:00:00',5494.78); 

现在的问题是，如何更新基础上，在表中的排名列“来自支付组的总额（金额）按客户编号排序”？

Answer 1

编辑（删除SQL Server代码，SQLite不支持联接在UPDATE语句）：

我认为这应该是确定：

drop table if exists ranks; 

create temp table ranks (
    rank integer primary key, 
    customerNumber real, 
    total real 
); 

insert into ranks 
select null, customerNumber, total(amount) as total 
from payments 
group by customerNumber 
order by total desc; 


update payments set 
    rank = (
     select rank from ranks 
     where customerNumber = payments.customerNumber 
    ); 

select * from payments; 

补充：

这是非常非常肮脏的方式来做到这一点，它使用非常具体的子查询t o排名（这是某种SQLite行号）每个客户。正如我所说，很多很多sybqueries，但这是让它在一个查询中工作而不使用更新中的连接的代价。

update payments set 
    rank = (
     select 
      (select count(0) from (select total(amount) as t, customerNumber as c 
        from payments group by customerNumber order by t desc) t1 
      where t1.t >= t2.t) as rank 
     from (select total(amount) as t, customerNumber as c from payments group by customerNumber order by t desc) 
     t2 where t2.c = payments.customerNumber order by t2.t desc 
    );