Postgresql函数如果条件满足，则运行query else stop

Question

我需要一些关于以下postgres功能的帮助

我有下表和列：

array, array_length

我最初在中有几个数组，然后运行一个查询（实际上是一组3个查询），选择数组，附加它们，然后将附加的数组插入表中。

我需要循环这个插入查询，直到其中一个数组（select max（array_length）from table）达到预定义的大小，比如长度50。

我基本上需要写一些

begin 
for (select max(array_length) from table)<50
loop 
(drop table if exists super_item_temp;
ALTER TABLE super_item
    RENAME TO super_item_temp;

create table super_item as 
select distinct * from super_item_temp;

insert into super_item 
select ...calculations... from super_item)
end

但是我找不到正确的语法来写这个

我http://www.postgresql.org/docs/8.4/static/plpgsql-control-structures.html看了手册，但找不到任何有用的东西。

任何提示或链接指向我正确的方向将不胜感激！谢谢你。

编辑：我试过了

创建或替换函数函数_name（）将int4返回为“声明r记录”；

开始时（从super_item中选择最大（array_length））

loop
drop table if exists super_item_temp;
    ALTER TABLE super_item
        RENAME TO super_item_temp;

    create table super_item as 
    select distinct * from super_item_temp;


    insert into super_item 
    select old_array, 
    array_sort_unique( array_agg(added_item) || a.old_array) as new_array, 
    array_length(array_sort_unique( array_agg(added_item) || a.old_array),1)

    from (
            select 
    a.new_array as old_array,

    case when string_to_array(b.item2::text, ",")::int[] <@ a.new_array and string_to_array(b.item1::text, ",")::int[] <@ a.new_array then null
     when string_to_array(b.item2::text, ",")::int[] <@ a.new_array then b.item1 else b.item2 end  as added_item

    from

    super_item a 
    left join pairs b 

    on  string_to_array(b.item1::text, ",")::int[] <@ a.new_array or  string_to_array(b.item2::text, ",")::int[] <@ a.new_array /**any item from pair is in array**/
    where 1=1

    group by a.new_array, 2

    having sum(b.count)>10
    and sum(b.offset)<=0 
    and 
    case 
    when string_to_array(b.item2::text, ",")::int[] <@ a.new_array and string_to_array(b.item1::text, ",")::int[] <@ a.new_array then null
    when string_to_array(b.item2::text, ",")::int[] <@ a.new_array then b.item1 else b.item2 end is not null
    /**new item is not null**/
    order by 2 desc
    )a
    group by 1;
END LOOP;

返回1； END；'LANGUAGE plpgsql；

选择function_name（）作为输出；

现在我得到一个函数参数的错误“unknown column”，它应该是一个带引号的分隔符。

错误：列“existiert nicht第18行：在字符串到数组（b.item1:：text，”，“”：：int[]

Answer 1

试试这个，在我的电脑上它编译得很好（但我不知道它是否能按预期工作）。

CREATE OR REPLACE FUNCTION append_super_item ()
RETURNS integer AS $length$
declare
    length integer;
BEGIN
  WHILE (select max(array_length) from super_item)<50
    LOOP
       drop table if exists super_item_temp;
       ALTER TABLE super_item
            RENAME TO super_item_temp;
       Create table super_item as 
         select distinct * from super_item_temp;
       insert into super_item 
    select old_array, 
    array_sort_unique( array_agg(added_item) || a.old_array) as new_array, 
    array_length(array_sort_unique( array_agg(added_item) || a.old_array),1)

    from (
            select 
    a.new_array as old_array,

    case when string_to_array(b.item2::text, ",")::int[] <@ a.new_array and string_to_array(b.item1::text, ",")::int[] <@ a.new_array then null
     when string_to_array(b.item2::text, ",")::int[] <@ a.new_array then b.item1 else b.item2 end  as added_item

    from

    super_item a 
    left join pairs b 

    on  string_to_array(b.item1::text, ",")::int[] <@ a.new_array or  string_to_array(b.item2::text, ",")::int[] <@ a.new_array /**any item from pair is in array**/
    where 1=1

    group by a.new_array, 2

    having sum(b.count)>10
    and sum(b.offset)<=0 
    and 
    case 
    when string_to_array(b.item2::text, ",")::int[] <@ a.new_array and string_to_array(b.item1::text, ",")::int[] <@ a.new_array then null
    when string_to_array(b.item2::text, ",")::int[] <@ a.new_array then b.item1 else b.item2 end is not null
    /**new item is not null**/
    order by 2 desc
    )a
    group by 1;
    end loop;
end $length$ LANGUAGE plpgsql;