redshift交错排序键的顺序是否重要？

Question

where子句是否以指定的顺序指定交错排序键是否重要？我想是的。如果订单很重要，为什么查询计划程序无法自行解决这个问题呢？

Answer 1

我认为订单并不重要。

在第34章Comparing Sort Styles＆＃34;中，我找到了一些有趣的基准。

表格方案：

create table cust_sales_date_interleaved as
(select c_custkey, c_nation, c_region, c_mktsegment, d_date::date, lo_revenue
from customer, lineorder, dwdate
where lo_custkey = c_custkey 
and lo_orderdate = dwdate.d_datekey
and lo_revenue > 0)
interleaved sortkey (c_custkey, c_region, c_mktsegment, d_date);

查询＃1

select max(lo_revenue), min(lo_revenue)
from cust_sales_date_interleaved
where c_region = 'ASIA'
and c_mktsegment = 'FURNITURE';

查询＃2

select max(lo_revenue), min(lo_revenue)
from cust_sales_date_interleaved
where d_date between '01/01/1996' and '01/14/1996'
and c_mktsegment = 'FURNITURE'
and c_region = 'ASIA';

我们有结果

查询＃1 - 1.46 s

查询＃2 - 0.80 s

查询＃2更快，因为它有d_date的额外过滤器。