CODEWITHSUNDEEP
mysqlselectunique
Aufwind
Aufwind

Reputation: 26278

How to select unique rows from one to many relationed tables in MySQL?

As MySQL-Noob I am struggeling to solve the following task: Assuming 3 tables: location groceries and person with the following properties:

  1. Every table has a primary autoincremented integer id and a varchar column.
  2. The entrys in the tables are all unique.
  3. Every entry in groceries has a foreign key of an entry in person.
  4. It is possible that more then one entry from groceries share the same foreign key.
  5. (3) and (4) apply to person and location, too

So we have a many to one to many relation. How can I select every triple (groceries_product, person_name, location_name) where person_name does not occur more then once?

Example:

tables:  groceries                  | person      | location
------------------------------------ ------------- -------------------------
columns: id  product      person_id | id  name    | id  name  person_id
------------------------------------ ------------- -------------------------
         1   hamburger    1         | 1   Peter   | 1   home  1
         2   cheeseburger 1         | 2   Tom     | 2   work  1
         3   carot        1         |             | 3   zoo   2 
         4   potatoe      1         |             |
         5   mango        2         |             |

All the triples you can create in which Peter occures are irrelevant. I want only triples like (mango, Tom, zoo), because Tom does occure only once in all possibilities. I hope my question ist understandable. :-)

Upvotes: 0

Views: 2719

Answers (2)

GreenTurtle
GreenTurtle

Reputation: 1262

I think you have to do a subselect to get your result:

SELECT groceries.product, person.name, location.name
FROM person
LEFT JOIN groceries ON person.id = groceries.person_id
LEFT JOIN location ON person.id = location.person_id
WHERE person.id
IN (
SELECT person.id
FROM person
LEFT JOIN groceries ON person.id = groceries.person_id
LEFT JOIN location ON person.id = location.person_id
GROUP BY person.id
HAVING count( person.id ) =1
)

Upvotes: 2

Bohemian
Bohemian

Reputation: 425258

select l.name, m.name, r.name
from `left` l
left join middle m on m.id = l.middle_id
left join `right` r on m.id = r.middle_id;

Upvotes: -1

Related Questions