cube关联后的distinct问题

[高扬02]

请教各位：

我需要关联两个fact表，所以目前是分别构建两个cube，要关联的字段为user_id，我将它设置同时为dimension和measure，以便可以取得明细（dimension）和计算count_distinct（measure），且不用开启query-push-down
但两个子查询分别命中了对应的cuboid，可是最外层去重查询报错，未知原因？

sql如下：
select count(distinct a.r_id) from
(select r.user_id as r_id,l.user_id as l_id from
(select user_id from register where data_day='2018-09-01' and 
app_id='403_20170811' )r
inner join
(select user_id from login where data_day='2018-09-01' and 
app_id='403_20170811')l
on l.user_id=r.user_id
)a;

报错：
Can't create EnumerableAggregate! while executing SQL: "select count(distinct 
a.r_bmid) from (select r.bmid_user as r_bmid,l.bmid_user as l_bmid from (select 
bmid_user from wl_register_bm_mes where data_day='2018-09-01' and 
app_id='403_20170811' )r inner join (select bmid_user from wl_login_new_mes 
where data_day='2018-10-14' and app_id='403_20170811')l on 
l.bmid_user=r.bmid_user )a LIMIT 50000"

期待解答。thx!