奇数“分组依据”使用Oracle时子查询出错

Question

简单的背景知识——我需要找到通过我们公司实验室的案件的终止率，并按案件类型和月份分组。到目前为止，我想出了这个：

SELECT BPI.TYPE,
       EXTRACT(MONTH FROM CS.RECEIVED_DATE) MONTH,
       COUNT(*) termed_cases
  FROM CELL_SOURCE cs
  JOIN BASIC_PATHOLOGY_INFO bpi ON CS.CELL_SOURCE_ID = BPI.CELL_SOURCE_ID
  JOIN RECENT_CELL_SOURCE_STATUS rcss ON CS.CELL_SOURCE_ID = RCSS.CELL_SOURCE_ID
 WHERE type IS NOT NULL
  AND CS.RECEIVED_DATE > to_date('03/01/2011', 'MM/DD/YYYY/')
  AND RCSS.STATUS like 'Term%'
GROUP BY BPI.TYPE, EXTRACT(MONTH FROM CS.RECEIVED_DATE)
ORDER BY month, type

这很容易找到所有已被命名的案例。然而，当我想找到汇率时，我遇到了一些问题。我尝试使用子查询来捕获每种类型的案例总数，无论其状态如何，如下所示：

COUNT(*)/(SELECT COUNT(*) 
            FROM CELL_SOURCE cs_1 
            JOIN BASIC_PATHOLOGY_INFO bpi_1 ON CS_1.CELL_SOURCE_ID = BPI_1.CELL_SOURCE_ID
          WHERE BPI_1.TYPE = BPI.TYPE 
            AND EXTRACT(month from CS_1.RECEIVED_DATE) = EXTRACT(MONTH FROM CS.RECEIVED_DATE)) termed_cases

但是，这会引发 ORA-00979: not a GROUP BY expression 错误，并突出显示 <来自子查询的 code>BPI.TYPE。

有人知道我的错误实际上可能是什么吗？另外，分析函数在这里会比聚合函数更好吗？

Answer 1

因此，您需要两个计数：当月所有案例的总数和仅术语案例的总数。最简单的方法是使用 CASE() 函数执行条件计数，如下所示：

SELECT BPI.TYPE,
       EXTRACT(MONTH FROM CS.RECEIVED_DATE) MONTH,
       COUNT(*) all_cases,
       sum(case when RCSS.STATUS like 'Term%' then 1 else 0 end ) termed_cases
  FROM CELL_SOURCE cs
  JOIN BASIC_PATHOLOGY_INFO bpi ON CS.CELL_SOURCE_ID = BPI.CELL_SOURCE_ID
  JOIN RECENT_CELL_SOURCE_STATUS rcss ON CS.CELL_SOURCE_ID = RCSS.CELL_SOURCE_ID
 WHERE tumor_type IS NOT NULL
  AND CS.RECEIVED_DATE > to_date('03/01/2011', 'MM/DD/YYYY/')
GROUP BY BPI.TUMOR_TYPE, EXTRACT(MONTH FROM CS.RECEIVED_DATE)
ORDER BY month, tumor_type

请注意，我已从 WHERE 子句中删除了 LIKE 过滤器。

Answer 2

select bpi.type , month, 
       termed_cases / 
   (select count(*) 
    from CELL_SOURCE cs_1 
    inner join BASIC_PATHOLOGY_INFO bpi_1 
        on CS_1.CELL_SOURCE_ID = BPI_1.CELL_SOURCE_ID
    where BPI_1.TUMOR_TYPE = BPI.TUMOR_TYPE 
    and extract(month from CS_1.RECEIVED_DATE) = extract(MONTH FROM CS.RECEIVED_DATE)
    )
from (
select BPI.TYPE,
extract(MONTH FROM CS.RECEIVED_DATE) MONTH,
count(*) termed_cases
from CELL_SOURCE cs
inner join BASIC_PATHOLOGY_INFO bpi
    on CS.CELL_SOURCE_ID = BPI.CELL_SOURCE_ID
inner join RECENT_CELL_SOURCE_STATUS rcss
    on CS.CELL_SOURCE_ID = RCSS.CELL_SOURCE_ID
where tumor_type is not null
and CS.RECEIVED_DATE > to_date('03/01/2011', 'MM/DD/YYYY/')
and RCSS.STATUS like 'Term%'
group by BPI.TYPE, extract(MONTH FROM CS.RECEIVED_DATE)
)
order by month, type