使用“分组依据”一次更新多个非重复行

Question

我听从了Generate a random number of non duplicated random number in [0, 1001] through a loop 的帮助。

但我不能将其应用于我的案例。当我这样做时：

update weighed_directed_edge set endpoint= trunc(1000 * random()+ 1)
from  generate_series(1,10) group by 1 where startpoint= from_point;

更新它抱怨的endpointId：

ERROR:  syntax error at or near "group"
LINE 1: ...nc(1000 * random()+ 1) from generate_series(1,10) group by 1.

我也试过了：

insert into weighed_directed_edge (startpoint,endpoint)
values (from_point, trunc((1000 * random()+ 1) )
FROM generate_series(1, directed2number)
GROUP  BY 1 ;

insert into weighed_directed_edge (startpoint, endpoint, costs)
select 1, trunc(1000 * random()+ 1) from  generate_series(1,10) group by 1,1;

不工作。我需要从另一个表中设置的相同点中选择起点和终点，以填充weighed_directed_edge 表。

• cross_point 表：有 1001 个交叉点。

                               Table "public.cross_point"
      Column    |  Type   |                                   Modifiers                              
  --------------+---------+---------------------------------------------------------------------
   intersection | integer | not null default nextval  ('intersection_intersection_seq'::regclass)
   x_axis       | integer | not null
   y_axis       | integer | not null
  Indexes:
      "intersection_pkey" PRIMARY KEY, btree (intersection)
      "intersection_x_axis_key" UNIQUE, btree (x_axis, y_axis)
  

• weighed_directed_edge 表：

         Table "public.weighed_directed_edge"
         Column   |       Type       | Modifiers 
      ------------+------------------+-----------
       startpoint | integer          | 
       endpoint   | integer          | 
       costs      | double precision | 
      Indexes:
          "weighed_directed_edge_startpoint_key" UNIQUE, btree (startpoint, endpoint)
      Foreign-key constraints:
          "weighed_directed_edge_endpoint_fkey" FOREIGN KEY (endpoint) REFERENCES cross_point(intersection)
          "weighed_directed_edge_startpoint_fkey" FOREIGN KEY (startpoint) REFERENCES cross_point(intersection)
  

• 每个起点（依次从点数列中获取）对应一个随机数的端点（从 1001 个点列中随机获取）。 端点数量在[1,7]中。

• 费用将基于起点和起点之间的距离 端点。

要求：

• 起点和终点的组合应该是唯一的。

• 起点集包含表cross_point的所有1001个交集

• 端点都来自同一个交点集。

• 每个星点最多有 7 个端点匹配，（最多 7 个道路交叉口），并且 端点匹配的数量是随机选择的。

我在下面的代码中跳过了成本的计算：

--The largest number of intersection chosen is 7, but this could be altered
create or replace function popluate_weighed_directed_edge() returns void as $$
declare
from_point integer;
to_point integer;
directed2number integer; --the number of node this startpoint leads to
counter integer;
factor float; 
weight numeric;
start_pointer record;

begin
for start_pointer in select * from cross_point
loop
from_point := start_pointer.intersection;
    directed2number := trunc(Random()*7+1);
    counter := directed2number;
        while counter > 0
        loop
insert into weighed_directed_edge (startpoint) select from_point from  generate_series(1,10) ;
update weighed_directed_edge set endpoint= trunc(1000 * random()+ 1) from  generate_series(1,10) group by 1 where startpoint= from_point ;
update weighed_directed_edge set costs= trunc(1000 * random()+ 1) from  generate_series(1,10) group by 1 where startpoint= from_point ;
        counter := counter - 1;
        end loop;
end loop;
end
$$ language plpgsql;

Answer 1

在循环中，您可以绘制一个随机数并仅在它不存在时插入（并减少循环计数器）。 伪代码：

while (counter > 6) 
loop:
  this = 1+ random() *1000
  insert into weighed_directed_edge (startpoint, endpoint, costs)
  VALUES ( :frompoint, :this, xxx* random() )
  WHERE NOT EXISTS (
    SELECT(*) FROM weighed_directed_edge nx
    WHERE nx.startpoint = :frompoint
    AND nx.endpoint = :this
    );

  if (rowcount > 0) counter -= 1;
end loop;