Greenplum或DeepGreen数据库查看表倾斜的方法总结(3)

前几天写的两篇博客介绍如何根据表的数据文件来查看数据倾斜度，如下：

http://blog.csdn.net/jiangshouzhuang/article/details/51792580

http://blog.csdn.net/jiangshouzhuang/article/details/51850975

下面总结其他的一些常用方法：

1、使用隐藏字段gp_segment_id
select gp_segment_id ,count(1) from test group by 1 order by 1;


2、使用get_ao_distribution
select * from get_ao_distribution('test');


3、使用 pg_relation_size和get_dist_random
select gp_segment_id,pg_relation_size(oid)
  from gp_dist_random('pg_class')
 where relname = 'test';
 
4、SQL查询表发生比较严重的数据倾斜：
select tabname ,max(size)/(avg(size) + 0.001) as max_div_avg, sum(size) total_size
from (
    select gp_segment_id, oid::regclass tabname ,pg_relation_size(oid) size
      from gp_dist_random('pg_class')
     where relkind = 'r' and relstorage in ('a','h')
)t group by tabname
   order by max_div_avg desc;