INNER JOIN可以用同一个表中的另一个值替换列的值 - 对于所有行

Question

我需要编写一个查询来将列中的一个值替换为同一个表中另一行的值。 我试图使用一个简单的内连接，但我没有看到我需要的结果。

这是BANK_LOANS表的一个例子：

MONTH_ID   LOAN_NUMBER    BANK_NAME       AMOUNT
========   ===========    ==========      ========
76         00-100         Bank One        100.00
75         00-100         The Bank One    150.00
74         00-100         Bank 1          150.00
76         00-200         Another Bank    300.00 
75         00-200         Another Bank    500.00

数据是从外部来源加载的，不会验证银行名称，因此给定贷款号码的银行名称的拼写可能会有所不同。 客户希望BANK_NAME在数据库中的任何内容，最大MONTH_ID用于该LOAN_NUMBER的所有结果，并且他们希望结果按BANK_NAME排序。 因此，例如，贷款00-100需要所有行的BANK_NAME为“Bank One”。

我尝试使用内部联接来执行此操作，但未获得正确的结果

SELECT name.BANK_NAME, 
       bl.LOAN_NUMBER,
       bl.AMOUNT
FROM BANK_LOANS    
INNER JOIN BANK_LOANS home
    ON name.LOAN_NUMBER = bl.LOAN_NUMBER
    AND name.MONTH_ID = 67  --the max id provided to the query
ORDER BY name.BANK_NAME, bl.LOAN_NUMBER, bl.MONTH_ID DESC

我想我有一个子查询，但它有点难看。 我想知道是否有更好的方法（最佳实践）通过连接或其他oracle函数来实现这一点。

这似乎有效，但它只是错了：

SELECT name.BANK_NAME, 
       bl.LOAN_NUMBER,
       bl.AMOUNT
FROM BANK_LOANS, 
     (SELECT bl2.BANK_NAME, bl2.LOAN_NUMBER
      FROM BANK_LOANS
      WHERE bl2.MONTH_ID = 76  --max month id provided to query
     ) name
WHERE name.LOAN_NUMBER = bl.LOAN_NUMBER
ORDER BY name.BANK_NAME, bl.LOAN_NUMBER, bl.MONTH_ID DESC

Answer 1

with
     test_data ( month_id, loan_number, bank_name, amount ) as (
       select 76, '00-100', 'Bank One'    , 100.00 from dual union all
       select 75, '00-100', 'The Bank One', 150.00 from dual union all
       select 74, '00-100', 'Bank 1'      , 150.00 from dual union all
       select 76, '00-200', 'Another Bank', 300.00 from dual union all
       select 75, '00-200', 'Another Bank', 500.00 from dual
     )
-- end of test data (not part of the solution).
-- SQL query begins BELOW THIS LINE; use with your table.
select month_id, loan_number, 
       first_value(bank_name) 
              over (partition by loan_number order by month_id desc) as bank_name,
       amount
from   test_data
;

MONTH_ID LOAN_N BANK_NAME    AMOUNT
-------- ------ ------------ ------
      74 00-100 Bank One        150
      75 00-100 Bank One        150
      76 00-100 Bank One        100
      75 00-200 Another Bank    500
      76 00-200 Another Bank    300

5 rows selected.

Answer 2

使用分析函数：

select bl.loannumber, bl.amount,
       max(bank_name) keep (dense_rank first order by month_id desc) over (partition by loannumber) as bank_name
from bank_loans bl;

Answer 3

SELECT month_id,
   loan_number,
   MIN(bank_name) over(PARTITION BY loan_number ORDER BY month_id DESC),
   amount
FROM bank_loans;