array_ The contains() function uses

One 、array_contains function ：

Hive Medium array_contains Function and SQL Medium in keyword The operation is similar to , Used to determine the inclusion （array_contains） Or does not contain （!array_contains） Relationship . And in The difference is array_contains It can be used to judge the same id Whether the same field in multiple records of contains one or more specified values . Pay attention to the consistency of field types , If inconsistent, you need to cast .

grammar ：
array_contains( Array , value ), Returns a Boolean value .

Two ,array_contains() Function in where Case in condition
Solve the problem of using too many conditions in a field , The query condition is invalid

select * 
from login 
where dt='20130101' 
  and (ver !='3.1' or ver !='3.2' 
                   or ver != '4.0' 
                   or ver != '5.2');

select * 
from login 
where dt='20130101' 
and (ver !='3.1' and ver !='3.2' 
                 and ver != '4.0' 
                 and ver != '5.2');

The purpose of this statement is to login from a user table , Inquire about 20130101 same day , The version is not 3.1,3.2,4.0,5.2 Login user record . however , stay hive0.90 perform , The last thing to come back is 20130101 All login user records of the day , The following query conditions seem to be ignored . Here, it seems that the same field does this kind of multiple != perhaps = The conditions of inquiry , Too much will fail .

So I modified the sentence , as follows :

select * 
from login 
where dt=20130101 
and !array_contains(split('3.1,3.2,4.0,5.2',','),ver);

3、 ... and 、array_contians() Direct application cases ：

Use array_contians() Label the monitoring indicators , Judge the alarm level .

select
    task_id,
    task_code,
    task_version,
    case
        when array_contains(collect_set(tag_id), 599)
            or array_contains(collect_set(tag_id), 1225)
        then 'L1- Level '
        when array_contains(collect_set(tag_id), 600)
            or array_contains(collect_set(tag_id), 1226)
        then 'L2- Level '
        when array_contains(collect_set(tag_id), 601)
            or array_contains(collect_set(tag_id), 1240)
        then 'L3- Level '
        when array_contains(collect_set(tag_id), 1166)
            or array_contains(collect_set(tag_id), 1241)
        then ' Eagle eye monitoring summary '
        else ' Non eagle eye monitoring task '
    end as core_tag_name
from
    fdm.fdm_buffalo_uniontask_tag_theme_info_new
where
    dt = sysdate( - 1)
group by
    task_id,
    task_code,
    task_version

Four 、 Use array_contians() Tag users , In user portrait , Play a role in dimension construction

select
   user_id,
   count(*) as card_number    --  Number of cards used 
   --  As long as any card is valid, it will be judged as VIP It works ,
   case when array_contains(collect_set(status_code),cast(1 as smallint)) 
        then 1 
        else 0 
    end AS effective_flag   --  Card valid identification 
from edw_users.dwd_edw_user_vipcard_df  --  user VIP Use the full scale for card purchase 
where dt = '${dt}'
and user_id > 0
and deleted_flag = 'N'
group by user_id