蜂巢查询 - 三个加盟与操作条件或连接两个表

Question

我面临着一个错误

“FAILED: Error in semantic analysis: Line 1:101 OR not supported in JOIN currently dob”

在运行下面提到的查询..

Insert Overwrite Local Directory './Insurance_Risk/Merged_Data' Select f.name,s.age,f.gender,f.loc,f.marital_status,f.habits1,f.habits2,s.employement_status,s.occupation_class,s.occupation_subclass,s.occupation from sample_member_detail s Join fb_member_detail f 
On s.email=f.email or 
s.dob=f.dob 
or (f.name=s.name and f.loc = s.loc and f.occupation=s.occupation) 
where s.email is not null and f.email is not null; 

谁能告诉我认为，在蜂房“OR”运营商可以使用或不？ 如果不是，那么应该是什么样的查询将给出与上述查询给出的结果相同的结果。 我有2个表，我想加入两个表中的任何一个与运算符的三个条件。 请帮忙..

Answer 1

对不起Hive只支持equi连接。你可以尝试从（你必须在非严格模式）这些表的全笛卡尔积选择：

Select f.name,s.age,f.gender,f.loc,f.marital_status,f.habits1,f.habits2,s.employement_status,s.occupation_class,s.occupation_subclass,s.occupation 
from sample_member_detail s join fb_member_detail f 
where (s.email=f.email 
or s.dob=f.dob 
or (f.name=s.name and f.loc = s.loc and f.occupation=s.occupation)) 
and s.email is not null and f.email is not null; 

Answer 2

你也可以使用UNION来得到相同的结果：

INSERT OVERWRITE LOCAL DIRECTORY './Insurance_Risk/Merged_Data' 
-- You can only UNION on subqueries 
SELECT * FROM (
    SELECT f.name, 
     s.age, 
     f.gender, 
     f.loc, 
     f.marital_status, 
     f.habits1, 
     f.habits2, 
     s.employement_status, 
     s.occupation_class, 
     s.occupation_subclass, 
     s.occupation 
    FROM sample_member_detail s 
    JOIN fb_member_detail f 
    ON s.email=f.email 
    WHERE s.email IS NOT NULL AND f.email IS NOT NULL; 

    UNION 

    SELECT f.name, 
     s.age, 
     f.gender, 
     f.loc, 
     f.marital_status, 
     f.habits1, 
     f.habits2, 
     s.employement_status, 
     s.occupation_class, 
     s.occupation_subclass, 
     s.occupation 
    FROM sample_member_detail s 
    JOIN fb_member_detail f 
    ON s.dob=f.dob 
    WHERE s.email IS NOT NULL AND f.email IS NOT NULL; 

    UNION 

    SELECT f.name, 
     s.age, 
     f.gender, 
     f.loc, 
     f.marital_status, 
     f.habits1, 
     f.habits2, 
     s.employement_status, 
     s.occupation_class, 
     s.occupation_subclass, 
     s.occupation 
    FROM sample_member_detail s 
    JOIN fb_member_detail f 
    ON f.name=s.name AND f.loc = s.loc AND f.occupation=s.occupation 
    WHERE s.email IS NOT NULL AND f.email IS NOT NULL; 

) subquery;