如何在 Oracle 10g 中获取不在 GROUP BY 子句中的其他列？

Question

我正在输入 SELECT 语句来检索每个客户第一次出现在销售收据历史记录中的信息，我有以下语句来检索有关 client_id 和 [=22 的信息=] 日期，但我需要 sales_receipt_id 每张收据。你能帮助我吗？谢谢！

我正在使用 Toad for Oracle 9.7.2.5 中连接的 Oracle 数据库 10G。

SELECT cli_codigo AS client_id, MIN(cmp_fecha_contable) AS sales_receipt_date
FROM   comprobantes
WHERE  cli_codigo IS NOT NULL
GROUP BY cli_codigo
ORDER BY cli_codigo ASC;

这句话实际上我们运行很好，但我也需要 receipt_id。（所有列都在同一个 table 中，名为 "comprobantes"。

我期望输出：

RECEIPT_ID  CLIENT_ID   RECEIPT_DATE    
201203000174    061     3/19/2012    
201203000027    118     3/19/2012    
201203000028    166     3/19/2012    
201203000029    139     3/19/2012    
201203000031    055     3/19/2012

Answer 1

select comprobantes.* 
from comprobantes 
left join 
(
SELECT cli_codigo AS client_id, MIN(cmp_fecha_contable) AS sales_receipt_date
FROM   comprobantes
WHERE  cli_codigo IS NOT NULL
GROUP BY cli_codigo
ORDER BY cli_codigo ASC
) a on a.client_id = comprobantes.cli_codigo and comprobantes.cmp_fecha_contable = a.sales_receipt_date

Answer 2

列必须在分组依据或聚合函数中。所以你可以这样做：

SELECT cli_codigo AS client_id, MIN(cmp_fecha_contable) AS sales_receipt_date, min(RECEIPT_ID) 
FROM   comprobantes
WHERE  cli_codigo IS NOT NULL
GROUP BY cli_codigo
ORDER BY cli_codigo ASC

或者

SELECT cli_codigo AS client_id, MIN(cmp_fecha_contable) AS sales_receipt_date, RECEIPT_ID
FROM   comprobantes
WHERE  cli_codigo IS NOT NULL
GROUP BY cli_codigo, RECEIPT_ID
ORDER BY cli_codigo ASC;

Answer 3

您可以使用 keep:

SELECT cli_codigo AS client_id,
       MIN(cmp_fecha_contable) AS sales_receipt_date,
       MIN(receipt_id) KEEP (DENSE_RANK FIRST ORDER BY cmp_fecha_contable) as first_receipt_id
FROM comprobantes
WHERE cli_codigo IS NOT NULL
GROUP BY cli_codigo
ORDER BY cli_codigo ASC;

这本质上是一个聚合 first_value() 函数。虽然语法有点笨拙，但 keep 在聚合查询中是一个非常强大的结构。

如何在 Oracle 10g 中获取不在 GROUP BY 子句中的其他列？

How to get other columns with are not in GROUP BY clause in Oracle 10g?

sql

oracle

group-by

oracle10g