Question

I have the following SQL:

SELECT *
FROM [Database].dbo.[TagsPerItem] 
INNER JOIN [Database].dbo.[Tag] ON [Tag].Id = [TagsPerItem].TagId 
WHERE [Tag].Name IN ('home', 'car') 

and it returns:

Id TagId  ItemId ItemTable Id Name  SiteId
------------------------------------------
1  1      1      Content   1  home  1
2  1      2      Content   1  home  1
3  1      3      Content   1  home  1
4  2      4      Content   2  car   1
5  2      5      Content   2  car   1
6  2      12     Content   2  car   1

instead of just two records, which these names are “home” and “car”. How can I fix it?

Thanks.

Answer 1

There are two alternatives – using JOINs:

SELECT *
  FROM [Database].dbo.[TagsPerItem] tpi
  JOIN [Database].dbo.[Tag] t_home ON t_home.id = tpi.tagid
                                  AND t_home.name = 'home'
  JOIN [Database].dbo.[Tag] t_car ON t_car.id = tpi.tagid
                                 AND t_car.name = 'car'

…or HAVING COUNT(DISTINCT t.name) = 2:

  SELECT --column list - fill with specific appropriate columns
    FROM [Database].dbo.[TagsPerItem] tpi
    JOIN [Database].dbo.[Tag] t ON t.id = tpi.tagid
                               AND t.name IN ('home', 'car')
GROUP BY --column list - match columns declarations with the SELECT list without aggregates used on them
  HAVING COUNT(DISTINCT t.name) = 2

• The COUNT(DISTINCT is necessary, otherwise two relations of “car” would be a false positive (assuming possible in the data model)
• Not all databases support COUNT(DISTINCT...); while tedious to construct, JOINing to the same table multiple times is a safer method