Redshift - 无法将查询结果加载到表中 - 领导节点问题

Question

我的目标是将查询结果加载到表中，以存储给定表的列名的有序列表。 （然后，我将使用将传递给动态 sql 的 listagg 函数将所有这些列名填充到单个列中。）我无法将其加载到表中的原因是因为系统表查询在领导节点上计算，但此查询不在领导节点上执行，也没有办法强制它在领导节点上执行。任何想法如何让它成功执行？

create table temp_columns
as
select
cast(t1.columnname as varchar) as columname
--,
--cast(t2.ordinal_position as integer) as ordinal_position
FROM 
(

SELECT 

        cast(schemaname as varchar) as schemaname,
        cast(tablename as  varchar) as tablename,
        cast("column" as varchar) as columnname

FROM PG_TABLE_DEF
WHERE 
schemaname = 'schema1'
and tablename = 'table1'
)
t1
 join information_schema.columns t2
    on t1.schemaname = t2.table_schema 
    and t1.tablename = t2.table_name
    and t1.columnname = t2.column_name
WHERE 
t1.schemaname = 'schema1'
and t1.tablename = 'table1'
order by t2.ordinal_position; 

Answer 1

我们通过先创建表然后执行 INSERT INTO 来解决这个问题。

类似：

create table public.temp_columns
(
columname varchar(255),
ordinal_position int
);

insert into public.temp_columns
select
cast(t1.columnname as varchar) as columname,
cast(t2.ordinal_position as integer) as ordinal_position
FROM 
(

SELECT 

        cast(schemaname as varchar) as schemaname,
        cast(tablename as  varchar) as tablename,
        cast("column" as varchar) as columnname

FROM PG_TABLE_DEF
WHERE 
schemaname = 'schema1'
and tablename = 'table1'
)
t1
 join information_schema.columns t2
    on t1.schemaname = t2.table_schema 
    and t1.tablename = t2.table_name
    and t1.columnname = t2.column_name
WHERE 
t1.schemaname = 'schema1'
and t1.tablename = 'table1'
order by t2.ordinal_position;