Question

我有以下方式的数据：

ORDER_NO  FULFILL_ID  SOURCE_LOC   ITEM 
  100                    11001     0021
  100                    11001     0031
  100                    12001     0014
  100                    13001     0053

每当source_loc相同时，它应该给出相同的fulfill_id。 ID应从1开始并递增1 ..因此数据应更新为

ORDER_NO  FULFILL_ID  SOURCE_LOC   ITEM 
  100         1           11001     0021
  100         1           11001     0031
  100         2           12001     0014
  100         3           13001     0053

如何更新列？

Answer 1

这可以使用单个MERGE语句完成：

merge into orders o
using
(
  select order_no, 
         dense_rank() over (order by source_loc) as rn, 
         source_loc, 
         item
  from orders
) t on (o.order_no = t.order_no and o.item = t.item and o.source_loc = t.source_loc)
when matched then update 
  set fullfill_id = rn;

问题是：如果可以在运行时轻松计算，为什么需要保存此信息。

您可以创建如下视图：

create or replace view orders_2
as
select order_no, 
       dense_rank() over (order by source_loc) as fullfill_id, 
       source_loc, 
       item
from orders

Answer 2

我不知道这是否重要，但是如果新记录带有相同的{% extends "base.html" %} {% load render_table from django_tables2 %} {% load crispy_forms_tags %} {% block my_table %} <div> <form action="" method="get"> {% crispy filter.form filter.form.helper %} </form> </div> {% render_table table %} {% endblock %}字段，并且您希望确保它们保留并继承先前为source_loc分配的值，那么像这样可能会确保先前的ID被继承并分配新的ID：

fulfill_id

以这种方式，一旦将fulfill_id分配给source_loc，它就会被保留。像这样开始的数据：

update mytable t1
set t1.fulfill_id = (
  with already as (
    select distinct source_loc, fulfill_id
    from mytable
    where fulfill_id is not null
  ),
  final_id as (
    select distinct
      m.source_loc,
      coalesce (a.fulfill_id, dense_rank() over (order by m.source_loc)) as fulfill_id    
    from
      mytable m,
      already a
    where
      m.source_loc = a.source_loc (+)
  )
  select
    f.fulfill_id
  from final_id f
  where
    f.source_loc = t1.source_loc
)
where t1.fulfill_id is null;

会这样结束：

ORDER_NO  FULFILL_ID  SOURCE_LOC   ITEM 
  100         1           11001     0021
  100         1           11001     0031
  100         2           12001     0014
  100         3           13001     0053
  100                     12001     0014
  100                     10001     0014

Oracle：基于列值更新行

2 个答案: