在调度流程执行过程中发现流程的执行时间随着时间的增加，执行时间也在逐步变长，于是决定对流程的SQL进行优化处理：
原SQL如下：
SELECT
kuscode,
gstore_name,
from_unixtime(cast(store_open_date as BIGINT),'yyyy-MM-dd') store_open_date,
round(sum(if(member_ascribe_uscode = uscode,store_amount - tax_total_amount,0)),2) store_amount,
count(distinct if(member_ascribe_uscode = uscode,mid,'')) mid_count,
from tmp_order_sapply_g
where bi_dt >= '20220101'
and store_open_date >= date_sub(current_timestamp() ,37)
and store_open_date <> '9999-12-31 00:00:00'
and order_time < date_add(store_open_date,30)
GROUP BY uscode,store_name,store_open_date
优化后的SQL：
SELECT
kuscode,
gstore_name,
from_unixtime(cast(store_open_date as BIGINT),'yyyy-MM-dd') store_open_date,
round(sum(if(member_ascribe_uscode = uscode,store_amount - tax_total_amount,0)),2) store_amount,
count(distinct if(member_ascribe_uscode = uscode,mid,'')) mid_count,
from tmp_order_sapply_g
where bi_dt >= regexp_replace(cast(date_sub(current_timestamp() ,40) as STRING),'-','')
and store_open_date >= date_sub(current_timestamp() ,37)
and store_open_date <> '9999-12-31 00:00:00'
and order_time < date_add(store_open_date,30)
GROUP BY uscode,store_name,store_open_date
主要是针对分区键进行范围限定。
