Question

在此查询中：

update public.dctable
set dc = 1 
from (
        select d.*,
         min(time) over (partition by prod, customer, city, num) as mintime,
         row_number() over (partition by prod, customer, city, num order by 
time) as seqnum
  from public.dctable d
  ) dclk
  where dclk.seqnum <> 1

我收到错误：

错误：XX000：目标表必须是等值连词谓词的一部分

看不到我错过的东西。只需要设置dc = 1，其中时间＆gt; mintime和seqnum＆lt;＆gt; 1.

有什么想法吗？

谢谢！

Answer 1

您需要将源表连接到from表，如此

DF

Answer 2

我猜测Redshift遵循update上的Postgres语义。如果是这样的话：

update public.dctable d
    set dc = 1 
    from (select d.*,
                 min(time) over (partition by prod, customer, city, num) as mintime,
                 row_number() over (partition by prod, customer, city, num order by 
    time) as seqnum
          from public.dctable d
          ) d2
    where d2.prod = d.prod and d2.customer = d.customer and d2.city = d.city and
          d2.num = d.num and d2.mintime < d.time;

我不认为row_number()对你想做的事情是必要的。

Answer 3

您收到此错误是因为要更新的主表格未直接加入查询的using部分中的表格。 where子句必须将更新的表作为谓词的一部分。例如（使您的查询按原样运行）：

update public.dctable set dc = 1 
  from (
    select by prod, customer, city, num, min(time) as time
    from public.dctable group by 1,2,3,4
  ) dclk
where dclk.prod = dctable.prod
and dclk.customer =dctable.customer
and dclk.city=dctable.city
and dclk.num=dctable.num
and dclk.time=dctable.time

错误：目标表必须是equijoin谓词的一部分

3 个答案: