用postgresql特性简化group by 后取每组前n条记录的方法

前端之家收集整理的这篇文章主要介绍了用postgresql特性简化group by 后取每组前n条记录的方法前端之家小编觉得挺不错的,现在分享给大家,也给大家做个参考。

需求:

求一段时间内各生产线前3名坏机的原因及坏机数.


step 1,从将原始数据中抽取品质数据放于t96临时表

create temp table t96 
on commit drop 
as 
select * from t96_pd_log where recseq between '791G0' and '792G0' and f96_op='PD-QC'  ;
可以看下这个t96中对坏机原因的统计情况:
select f96_line,f96_result,sum(f96_qty) as qty from t96 group by f96_line,f96_result order by f96_line,qty desc



step2,一条sql就能得到需求的结果,其中用了with这个短句,类似于临时表或子查询的意思了,但如果用了with定义一个子集,在接着Select里面可以多处引用. 下面sql这个A子查询每条生产线用不良数倒序排,并每记录给一个行号 ...

with udc_qty_byline as 
(select f96_line,f96_result )
SELECT f96_line,qty FROM 
(
    SELECT *,ROW_NUMBER() OVER (PARTITION BY 
             f96_line order by qty DESC) AS Row_ID FROM udc_qty_byline
) AS A
WHERE Row_ID <=3 ORDER BY f96_line


Postgresql ispowerful.

猜你在找的Postgre SQL相关文章