如果联接的值在特定集中，则在JOIN查询中使用LIKE

Question

I have two sets:

1. dictionary of names
2. "names" extracted from emails

I want to find out if "names" from set 2. are similar/contains any of names in set 1 for example:

• johnsmith contains john so it will be ok.

How can I do this in HIVE SQL?

select a.*, b.name as name_from_email
from set2 a left join 
     set1 b 
     on a.email_name rlike concat('%',b.name,'%')

I got error: " Both left and right aliases encountered in JOIN ''%'':28:27, org.apache.hive.service.cli.operation.Operation:toSQLException:Operation.java:323, "

Answer 1

You can use instr like following:

on(1=1) where instr(a.email_name, b.name) != 0

I know this is painful but this will solve the problem

Cheers!!

Answer 2

You can rather painfully do this to get matches:

select a.*, b.name as name_from_email
from set2 a cross join 
     set1 b 
where a.email_name rlike concat('%', b.name, '%');

Getting the additional non-matched rows in is a pain. I think this should work:

select a.*, b.name as name_from_email
from set2 a cross join 
     set1 b 
where a.email_name rlike concat('%', b.name, '%');
union all
select a.*, null
from set2 a
where not exists (select 1
                  from set1 b
                  where a.email_name rlike concat('%', b.name, '%')
                 );

Note: You may find that this compiles but does not actually run.