total size of single column exceed the limit of hash join

我们目前已经更名为starrocks了哦。

另外咱们union的两个表大小是多少呢?

table1 : 215.455 GB
table2: 220.594 GB
table3: 227.260 GB

请问下有什么办法可以解决这个问题吗?

强制指定成Shuffle,并且增加下并行度试试

select product_id,big_label,sum(volume) volume from
((select product_id,
volume
from table1
where date >= ‘20220310’
and date <= ‘20220325’)
union all
(select product_id,
volume
from table2
where date >= ‘20220310’
and date <= ‘20220325’))
sale_product left join [shuffle] (select pid,big_label from table3) v5 on v5.pid = product_id where big_label = ‘baihou’
group by product_id,big_label
order by volume desc
limit 0,500

以前的版本,Join没有这个限制,可能结果已经错了

2.3版本,已经放开这个限制