【StarRocks版本】3.1.12 同一台FE,不同客户端地址发起查询,一台有数据,一个无数据

【StarRocks版本】3.1.12
【集群规模】3fe+3be(fe与be混部)
相同SQL,同一台FE,不同客户端地址发起查询,一台有数据,一个无数据,是什么原因?

查询有结果

查询无结果

这俩SQL看起来不一样啊

一样的,图一没有截全

有数据:

mysql> select count(1) from (
    -> select *
    -> from example_db.ads_guanxing_org_pro_sptm_sal_ttl t
    -> where 1 = 1
    ->   and is_noimport_store = 1
    ->   and last_update_date >= '2024-03-23'
    ->   and last_update_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    ->   and t.min_counter_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1,
    ->                                    10)
    ->   and is_first_spider_abnormal_date not between '2024-03-23' and substr(
    ->         cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    ->   and t.category2_no in ('50006843')
    ->   and t.view_amt >= '100.0'
    ->   and t.view_amt <= '500.0'
    ->   and arrays_overlap(split(style_name_list, '|'), split('小白鞋', ','))
    -> order by  view_amt desc) as a;
+----------+
| count(1) |
+----------+
|        4 |
+----------+
1 row in set (0.02 sec)

mysql> 

无数据:

MySQL [example_db]> select count(1) from (
    -> select *
    -> from example_db.ads_guanxing_org_pro_sptm_sal_ttl t
    -> where 1 = 1
    ->   and is_noimport_store = 1
    ->   and last_update_date >= '2024-03-23'
    ->   and last_update_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    ->   and t.min_counter_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1,
    ->                                    10)
    ->   and is_first_spider_abnormal_date not between '2024-03-23' and substr(
    ->         cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    ->   and t.category2_no in ('50006843')
    ->   and t.view_amt >= '100.0'
    ->   and t.view_amt <= '500.0'
    ->   and arrays_overlap(split(style_name_list, '|'), split('', ','))
    -> order by  view_amt desc) as a;
+----------+
| count(1) |
+----------+
|        0 |
+----------+
1 row in set (0.03 sec)

arrays_overlap(split(style_name_list, ‘|’), split(‘小白鞋’, ‘,’))

vs

arrays_overlap(split(style_name_list, ‘|’), split(’’, ‘,’))

饿,我知道了,SQL是我复制粘贴过去的,这句直接里面有中文,直接给我省略了

大佬,这两个SQL不能兼容吗?从2.5.8升级到3.1.12执行结果不一样
只有where条件的差异。

sql1:
mysql> select count(1) from (
    -> select
    -> *
    -> from
    -> example_db.ads_guanxing_org_pro_sptm_sal_ttl t
    -> where
    -> 1 = 1
    -> and is_noimport_store = 1
    -> and last_update_date >= '2024-03-23'
    -> and last_update_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    -> and t.min_counter_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1,10)
    -> and is_first_spider_abnormal_date not between '2024-03-23' and substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    -> and t.category2_no in ('50006843')
    -> and t.view_amt >= cast('100.0' as double)
    -> and t.view_amt <= cast('500.0' as double)
    -> and arrays_overlap(split(style_name_list,'|'),
    -> split('小白鞋',','))
    -> order by
    -> view_amt desc) as a ;
+----------+
| count(1) |
+----------+
|       10 |
+----------+
1 row in set (0.03 sec)
sql2:
mysql> select count(1) from (
    -> select
    -> *
    -> from
    -> example_db.ads_guanxing_org_pro_sptm_sal_ttl t
    -> where
    -> 1 = 1
    -> and is_noimport_store = 1
    -> and last_update_date >= '2024-03-23'
    -> and last_update_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    -> and t.min_counter_date <= substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1,10)
    -> and is_first_spider_abnormal_date not between '2024-03-23' and substr(cast(date_sub(from_unixtime(unix_timestamp()), 1) as string), 1, 10)
    -> and t.category2_no in ('50006843')
    ->   and t.view_amt >= '100.0'
    ->   and t.view_amt <= '500.0'
    -> and arrays_overlap(split(style_name_list,'|'),
    -> split('小白鞋',','))
    -> order by
    -> view_amt desc) as a ;
+----------+
| count(1) |
+----------+
|        4 |
+----------+
1 row in set (0.02 sec)

弄个profile看看