带有 postgres 窗口函数的重复行 [英] Duplicate lines with postgres window functions

问题描述

postgres 文档(http://www.postgresql.org/docs/9.1/static/tutorial-window.html) 讨论窗口函数.

The postgres documentation (http://www.postgresql.org/docs/9.1/static/tutorial-window.html) talks about window functions.

在他们的例子中:

SELECT salary, sum(salary) OVER (ORDER BY salary) FROM empsalary;

重复处理如下:

 salary |  sum  
--------+-------
   3500 |  3500
   3900 |  7400
   4200 | 11600
   4500 | 16100
   4800 | 25700 <-- notice duplicate rows have the same value
   4800 | 25700 <-- SAME AS ABOVE
   5000 | 30700
   5200 | 41100 <-- same as next line
   5200 | 41100 <--
   6000 | 47100
(10 rows)

你如何做同样的事情，使重复的​​行不会被赋予相同的值?换句话说，我希望这张表如下所示:

How do you do the same thing making it so that duplicate rows are NOT given the same value? In other words, I would like this table to look as follows:

 salary |  sum  
--------+-------
   3500 |  3500
   3900 |  7400
   4200 | 11600
   4500 | 16100
   4800 | 20900 <-- not the same as the next line
   4800 | 25700 <-- 
   5000 | 30700
   5200 | 35900 <-- not the same as the next line
   5200 | 41100 <--
   6000 | 47100
(10 rows)

推荐答案

在框架子句中使用 rows 而不是默认的 range

Use rows in the frame clause in instead of the default range

select
    salary,
    sum(salary) over (
        order by salary
        rows unbounded preceding
    )
from empsalary
;
 salary |  sum  
--------+-------
   3500 |  3500
   3900 |  7400
   4200 | 11600
   4500 | 16100
   4800 | 20900
   4800 | 25700
   5000 | 30700
   5200 | 35900
   5200 | 41100
   6000 | 47100

http://www.postgresql.org/docs/current/static/sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS

这篇关于带有 postgres 窗口函数的重复行的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！