Question

I didn’t even know how to come up with a good title, so I hope I can describe my problem in a right way 🙂

So I have a person table, and it has a N:N relationship with keywords via the table PersonKeywords.

Then I also have a Search table, and it also has a N:N relationship with keywords via the table SearchKeywords.

Now the person can have a relationship with keyword A and B, and the search record can have a relationship with the keywords A and C.

Now I want the person in my resultset, because it has at least one (in this ‘A’) of the keywords the search record has.

I also want the person who has ‘A’, the one with ‘C’, the one with ‘A’ and ‘C’, but not the one with only B.

So it’s a match on two lists, but I don’t know where to start to create such a statement…

Answer 1

So you have three people…

declare @persons table (id int identity(1,1), name varchar(10))
insert @persons (name) values ('Babs'),('Ken'),('Neville'),('Sue')

Babs has A and B, Ken has A and C, Neville has B only and Sue has C only

declare @personkeywords table (personid int, keyword varchar(5))
insert @personkeywords values (1,'a'),(1,'b'),(2,'a'),(2,'c'),(3,'b'),(4,'c')

The search is for A or C

declare @searchkeywords table (searchid int, keyword varchar(5))
insert @searchkeywords values (1,'a'),(1,'c')

So…

select distinct persons.*
from @persons persons
     inner join @personkeywords personkeywords on persons.id = personkeywords.personid
     inner join @searchkeywords searchkeywords on personkeywords.keyword = searchkeywords.keyword
where 
     searchkeywords.searchid = 1

Gives

1     Babs
2     Ken
4     Sue