1

目前正在进行 Oracle 到 T-SQL 语句的转换。这是我要转换的内容：

SELECT USC.USER_ID
  ,NVL2 (MAX(STREET_1), MAX(STREET_1) || CHR (10), '')
|| NVL2 (MAX(STREET_2), MAX(STREET_2) || CHR (10), '')
|| NVL2 (MAX(STREET_3), MAX(STREET_3) || CHR (10), '')
|| NVL2 (MAX(STREET_4), MAX(STREET_4) || CHR (10), '') AS STREETS
  ,MAX(ADDR.CITY) AS CITY
  ,MAX(ADDR.POSTAL_CODE) AS POSTAL_CODE
  ,MAX(ADDR.STATE_ID) AS STATE_ID
  ,MAX(ADDR.COUNTRY_ID) AS COUNTRY_ID
FROM ES_W_USER_CONT_INF USC
LEFT JOIN DW_ERSDB_LDSJOBS_ADDRESS ADDR ON (USC.ADDRESS_ID = 
ADDR.ADDRESS_ID)
GROUP BY USC.USER_ID

我最好的尝试。我意识到这很遥远，但我在编写案例陈述方面没有太多经验：

SELECT USC.USER_ID
,CASE 
 WHEN ADDR.STREET_1 IS NOT NULL THEN ADDR.STREET_1 || CHAR (10) 
 WHEN ADDR.STREET_1 IS NULL THEN '' 
 WHEN ADDR.STREET_2 IS NOT NULL THEN ADDR.STREET_2 || CHAR (10) 
 WHEN ADDR.STREET_2 IS NULL THEN '' 
 WHEN ADDR.STREET_3 IS NOT NULL THEN ADDR.STREET_3 || CHAR (10) 
 WHEN ADDR.STREET_3 IS NULL THEN '' 
 WHEN ADDR.STREET_4 IS NOT NULL THEN ADDR.STREET_4 || CHAR (10) 
 WHEN ADDR.STREET_4 IS NULL THEN '' 
 END
,MAX(ADDR.CITY) AS CITY
,MAX(ADDR.POSTAL_CODE) AS POSTAL_CODE
,MAX(ADDR.STATE_ID) AS STATE_ID
,MAX(ADDR.COUNTRY_ID) AS COUNTRY_ID
FROM DSS_ERS_STAGE.ES_W_USER_CONT_INF USC
LEFT JOIN DSS_ERS_STAGE.ES_W_ADDRESS ADDR ON (USC.ADDRESS_ID = 
ADDR.ADDRESS_ID)
GROUP BY USC.USER_ID
4

2 回答 2

2

您仍然有一些 Oracle 特定的项目，并且缺少一些有用的 SQL Server 功能。这是我的转换尝试：

 SELECT USC.USER_ID,
   COALESCE(ADDR.STREET_1 + CHAR (10), '') +
   COALESCE(ADDR.STREET_2 + CHAR (10), '') +
   COALESCE(ADDR.STREET_3 + CHAR (10), '') +
   COALESCE(ADDR.STREET_4 + CHAR (10), '') AS Streets,
  ,MAX(ADDR.CITY) AS CITY
  ,MAX(ADDR.POSTAL_CODE) AS POSTAL_CODE
  ,MAX(ADDR.STATE_ID) AS STATE_ID
  ,MAX(ADDR.COUNTRY_ID) AS COUNTRY_ID
FROM DSS_ERS_STAGE..ES_W_USER_CONT_INF USC
LEFT JOIN DSS_ERS_STAGE..ES_W_ADDRESS ADDR 
  ON USC.ADDRESS_ID = ADDR.ADDRESS_ID
GROUP BY USC.USER_ID

具体项目： COALESCE() 接受参数列表并返回第一个非空值。如果任何 ADDR.STREET_x 字段为空，则该字段和 CHAR(10) 的串联也将为空（除非您更改 ANSI_NULL 默认值）。

表名中的双点：我假设 DSS_ERS_STAGE 是数据库名称。如果在 SQL Server 中使用 Database Name.Table Name，则需要指定第三级 - 架构名称。在这种情况下，我假设默认模式名称 (dbo) 可以省略，但应包含分隔符。如果您使用不同的模式，例如“db2data”，您的表引用将如下所示：

FROM DSS_ERS_STAGE.db2data.ES_W_USER_CONT_INF USC

希望这对您有所帮助。

于 2018-06-01T20:33:42.943 回答
0

您应该能够将 NVL2 换成 COALESCE

SELECT USC.USER_ID
  ,COALESCE((MAX(STREET_1) + CHAR (10), '')
 + COALESCE(MAX(STREET_2) + CHAR (10), '')
 + COALESCE(MAX(STREET_3) + CHAR (10), '')
 + COALESCE(MAX(STREET_4) + CHAR (10), '') AS STREETS
  ,MAX(ADDR.CITY) AS CITY
  ,MAX(ADDR.POSTAL_CODE) AS POSTAL_CODE
  ,MAX(ADDR.STATE_ID) AS STATE_ID
  ,MAX(ADDR.COUNTRY_ID) AS COUNTRY_ID
FROM DSS_ERS_STAGE.ES_W_USER_CONT_INF USC
LEFT JOIN DW_ERSDB_LDSJOBS_ADDRESS ADDR ON (USC.ADDRESS_ID = 
ADDR.ADDRESS_ID)
GROUP BY USC.USER_ID

您可能必须连接 MAX(CHAR(10)) 我不确定。这不是功能问题，只是使用 GROUP BY 的语法。

于 2018-06-01T20:24:54.900 回答