CODEWITHSUNDEEP
mysql
randy
randy

Reputation: 1877

mysql trying to get a count with a one to many tables

I have tables like

products            product_attributes
==================  ========================================
| id | name      |  | id | product_id | attribute | value  |
==================  ========================================
| 1  | product 1 |  | 1  | 1          | min       | 2    |
| 2  | product 2 |  | 2  | 1          | max       | 5    |
| 3  | product 3 |  | 3  | 2          | min       | 10   |
| 3  | product 3 |  | 4  | 2          | max       | 15   |
| .. | ...       |  | 5  | 3          | min       | 1    |
==================  | 6  | 3          | max       | 100  |
                    | .. | ...        | ...       | ...  |
                    ========================================

I want to get all the products that fall within a range I can successfully get a list of the ids that fall with-in the range using

SELECT p.id 
  FROM `products` p INNER JOIN `product_attributes` AS pa ON p.id = pa.product_id   
  WHERE  pa.`min` <= 5 AND  pa.`max` >= 5 GROUP BY p.id

This gives me a list of the ids that i need. What i can not get to work is to get a TOTAL count of the ids instead of a list so i can not use the group by and then when i do that it messes up the count.

Is the only way to do this is by using the select above in a subselect and counting the results?

Thanks

Upvotes: 0

Views: 103

Answers (4)

fthiella
fthiella

Reputation: 49049

Given the structure above, where min and max are on different rows, I would use this:

select count(id) from (
  select p.id
  from
    `products` p INNER JOIN `product_attributes` AS pa
    ON p.id = pa.product_id
  where (pa.attribute='min' and pa.value>=5) OR
        (pa.attribute='max' and pa.value<=5)
  group by p.id
  having count(*)=2
) s

Upvotes: 1

Tebc
Tebc

Reputation: 121

Try a select count(*) from your select statement.

Upvotes: 0

Strawberry
Strawberry

Reputation: 33945

One approach is to apply the veneer of normalisation to the eav model before attempting to extract any information from it. In this case that might look like this:

 SELECT product_id
      , MAX(CASE WHEN attribute = 'min' THEN value END) `min` 
      , MAX(CASE WHEN attribute = 'max' THEN value END) `max` 
   FROM product_attributes
  GROUP   
     BY product_id;

 +------------+------+------+
 | product_id | min  | max  |
 +------------+------+------+
 |          1 |    2 |    5 |
 |          2 |   10 |   15 |
 |          3 |    1 |  100 |
 |        ... |  ... |  ... |
 +------------+------+------+

Now, what was your question again?

Upvotes: 0

hjpotter92
hjpotter92

Reputation: 80639

Just replace p.id with COUNT(1) in the query.

SELECT COUNT(1)
FROM `products` p
INNER JOIN `product_attributes` AS pa
    ON p.id = pa.product_id   
WHERE  pa.`min` <= 5
    AND  pa.`max` >= 5
GROUP BY p.id

Upvotes: 1

Related Questions