DB2 使用 SUM 函数导致 SQL 错误 [S1000]: 一般错误

Question

我正在尝试收集项目的销售数据，但我在主查询中用作 cte 的查询出现问题。

SELECT
    CONCAT(
      CONCAT(
       VARCHAR_FORMAT(INFO.DATE, 'YYYYMMDD'),
       CAST(INFO.INVOICE AS VARCHAR)
      ), 
      CAST(INFO.STORE AS VARCHAR)
    ) AS HEADER,
    CASE WHEN INFO.CUSTOMER_NUMBER IN (999990, 999991) 
      THEN 'CASH' ELSE 'CHARGE' END AS CUST_TYPE,
    INFO.STORE,
    INFO.INVOICE,
    INFO.DATE,
    INFO.SALES_UNITS,
    INFO.SALES
FROM
    DATALAKE.SLSINFO INFO
WHERE 
    INFO.DATE >= '2019-01-01'
    AND INFO.STORE = 163
    AND INFO.INVOICE = 100349
ORDER BY 
    HEADER ASC

此查询应该创建一个唯一的 header 然后使用客户编号将发票分类为现金客户或收费客户以及 return 一些相关值。

在这种情况下，我正在查看特定的 date/store/invoice 进行测试，结果是：

HEADER	CUST_TYPE	STORE	INVOICE	DATE	SALES_UNITS	SALES
20190101100349163	CASH	163	100349	2019-01-01	4	9.96
20190101100349163	CASH	163	100349	2019-01-01	1	10.99
20190101100349163	CASH	163	100349	2019-01-01	1	2.99

我试图通过添加 SALES_UNITS 和 SALES 列来为每个唯一 header 创建一行。我尝试使用的查询是这样的：

SELECT
    CONCAT(CONCAT(VARCHAR_FORMAT(INFO.DATE, 'YYYYMMDD'), CAST(INFO.INVOICE AS VARCHAR)), CAST(INFO.STORE AS VARCHAR)) AS HEADER,
    CASE WHEN INFO.CUSTOMER_NUMBER IN (999990, 999991) THEN 'CASH' ELSE 'CHARGE' END AS CUST_TYPE,
    INFO.STORE,
    INFO.INVOICE,
    INFO.DATE,
    SUM(INFO.SALES_UNITS) AS UNITS,
    SUM(INFO.SALES) AS SALES
FROM
    DATALAKE.SLSINFO INFO
WHERE 
    INFO.DATE >= '2019-01-01'
    AND INFO.STORE = 163
    AND INFO.INVOICE = 100349
ORDER BY 
    HEADER ASC

理想的输出应该是这样的

HEADER	CUST_TYPE	STORE	INVOICE	DATE	UNITS	SALES
20190101100349163	CASH	163	100349	2019-01-01	6	23.94

然而，第二个查询没有运行，我收到 SQL 错误 [S1000]：一般错误，没有其他信息。

我已经尝试在原始查询和第二个查询中按 header 分组，但它不允许我按任何分组，我还没有弄清楚为什么会这样。

任何见解将不胜感激！

Answer 1

您可以使用 with 子句为您的原始查询命名（我称之为 T，然后只需获取您的总和并按其他列分组。

      With T as (
  SELECT 
    CONCAT(
      CONCAT(
        VARCHAR_FORMAT(INFO.DATE, 'YYYYMMDD'), 
        CAST(INFO.INVOICE AS VARCHAR)
      ), 
      CAST(INFO.STORE AS VARCHAR)
    ) AS HEADER, 
    CASE WHEN INFO.CUSTOMER_NUMBER IN (999990, 999991) THEN 'CASH' ELSE 'CHARGE' END AS CUST_TYPE, 
    INFO.STORE, 
    INFO.INVOICE, 
    INFO.DATE, 
    INFO.SALES_UNITS, 
    INFO.SALES 
  FROM 
    DATALAKE.SLSINFO INFO 
  WHERE 
    INFO.DATE >= '2019-01-01' 
    AND INFO.STORE = 163 
    AND INFO.INVOICE = 100349
) 
select 
  Header, 
  cust_type, 
  store, 
  invoice, 
  date, 
  SUM(SALES_UNITS) AS UNITS, 
  SUM(SALES) AS SALES 
from 
  t 
group by 
  Header, 
  cust_type, 
  store, 
  invoice, 
  date 
order by 
  Header

DB2 使用 SUM 函数导致 SQL 错误 [S1000]: 一般错误

DB2 using SUM function causes SQL Error [S1000]: General error

sql

db2