Question

我有一个数据管道从Postgres中提取大量数据并将其推送到数据显示工具中。我的典型流程看起来像这样，并且工作正常：

create temporary table foo as 
select w.id, w.name, w.created_at
from waldo w;
\copy "foo" to './foo.txt';

然后我使用那个转出的foo.txt文件。（实际的查询要复杂得多，并且在十几个其他表上加入和分组，但你明白了。）

我开发了一个使用＆＃34; with query＆＃34;我无法弄清楚将结果导入文本文件的语法。此查询在我的交互式SQL窗口中正常运行：

WITH daze 
     AS (SELECT ( CURRENT_DATE - b.offs ) AS date 
         FROM   (SELECT Generate_series(0, CURRENT_DATE - '2015-01-15', 1) AS 
                        offs) 
                AS b) 
select w.id, w.name, w.created_at               
from waldo w, daze d
where d.date + 1 > w.created_at

但我需要将其添加到我的每日转储脚本中，并获取一个可以转储的临时表对象。而且这段代码没有运行：

create temporary table foo as 
WITH daze 
     AS (SELECT ( CURRENT_DATE - b.offs ) AS date 
         FROM   (SELECT Generate_series(0, CURRENT_DATE - '2015-01-15', 1) AS 
                        offs) 
                AS b) 
select w.id, w.name, w.created_at               
from waldo w, daze d
where d.date + 1 > w.created_at

我挣扎的语法是什么？提前谢谢。

Answer 1

抱歉，我不知道Postgres，只是TSQL，但你可以尝试在＃34; daze＆＃34;下面宣布另一个CTE。一个在同一个WITH里面？

此页面说明了Postgres语法中的可能性。 http://www.postgresql.org/docs/9.1/static/queries-with.html

WITH daze 
     AS (SELECT ( CURRENT_DATE - b.offs ) AS date 
         FROM   (SELECT Generate_series(0, CURRENT_DATE - '2015-01-15', 1) AS 
                        offs) 
                AS b)
    ,foo
     as (select w.id, w.name, w.created_at 
         from waldo w, daze d
         where d.date + 1 > w.created_at)
select * from foo

Answer 2

您的整个CTE是一种非常复杂的做法：

index.html

为什么不简单地说：

generate_series(CURRENT_DATE::timestamp, '2015-01-15'::timestamp, interval '-1 day') AS date

（请注意，CREATE TEMPORARY TABLE foo AS SELECT w.id, w.name, w.created_at FROM waldo w, generate_series((CURRENT_DATE + 1)::timestamp, '2015-01-16'::timestamp, interval '-1 day') AS d(date) WHERE d.date > w.created_at;条款中的+ 1已在WHERE来电中吸收。）

从查询＆＃34;＆＃34;的结果创建临时表

2 个答案: