Teradata 语法：unicode 分隔标识符

Question

我在尝试运行 Teradata 13.0 中的查询时遇到错误如何编写此查询？我将此 oracle 查询转换为 Teradata 查询：

select sb.name,
       sb.address1, 
       sb.address2, 
       sb.city, 
       sb.stateprovince,
       sb.postalcode, 
       sb.country, 
       sb.cid, 
       sb.item, 
       sb.mcnum, 
       sb.dmh, 
       sb.hy, 
       sb.firstname, 
       sb.lastname, 
       sb.email, 
       sb.monikerexec, 
       sb.res,
 (select cid from (select * from abc.submissions where res is not null order by res asc) where name = sb.name and item <> sb.item and res < sb.res and rownum =1) as mins,
  (select min(price) from ap.hist where name = sb.name) as minauc,
       sb.cat, 
       sb.uni
 from abc.submission sb
order by sb.item

错误：

Answer 1

听起来问题不在于缺少别名（这是一个问题），而是在子查询中需要 ORDER BY 以便适当的 cid可以在SELECT语句中的第一个子查询minauc中挑出

您可以使用 QUALIFY，而不是执行 ORDER BY 并尝试获取第一条记录，它使用窗口函数来选择您想要的记录：

SELECT 
    sb.NAME, 
    sb.address1, 
    sb.address2, 
    sb.city, 
    sb.stateprovince, 
    sb.postalcode,
    sb.country, 
    sb.cid, 
    sb.item, 
    sb.mcnum, 
    sb.dmh, 
    sb.hy, 
    sb.firstname, 
    sb.lastname, 
    sb.email, 
    sb.monikerexec, 
    sb.res,
    sb2.cid AS mins, 
    (
        SELECT min(price)
        FROM ap.hist
        WHERE NAME = sb.NAME
    ) AS minauc, 
    sb.cat, 
    sb.uni
FROM abc.submission sb
    LEFT OUTER JOIN abc.Submission sb2 ON
    sb.name = sb2.name AND
    sb.item <> sb.item AND
    sb2.res < sb.res AND
    sb2.res IS NOT NULL
QUALIFY ROW_NUMBER() OVER (PARTITION BY sb.name ORDER BY res asc) = 1

这使用 JOIN 将记录放在一起，然后，使用 QUALIFY 只允许通过具有最低 res 的连接记录。

Teradata 语法：unicode 分隔标识符

Teradata syntax: unicode delimited identifier

teradata