The database is not mine. The structure is
fid| uid | value
3 | 3 | spain
3 | 5 | France
2 | 3 | 45
6 | 3 | male
6 | 5 | female
2 | 5 | 32
The field ID is primary key in another table, I'd like to forget about. Im trying to find all uid that have values 'spain', 'male'
I have the following working.
SELECT uid
FROM DATABASE
WHERE value IN ('spain', 'male') GROUP BY uid HAVING COUNT(*) >= 2
The catch is as follows.. How would I select male, spain with value in range of 20-30 (that is males born in spain aged between 20-30?
Thanks a ton!!
You have a really messed up table. You are mixing unrelated things in value.
Try:
SELECT uid
FROM DATABASE a JOIN
DATABASE b USING (uid)
WHERE a.value IN ('spain', 'male')
AND b.value >= 20
AND b.value <= 30
GROUP BY uid
HAVING COUNT(*) >= 2
Note that I am comparing a string and an integer. You will need to test to see how well that works.
Also, I just put spain and male together, but perhaps they are really unrelated?
Does fid determine the type of value?
Try:
SELECT uid
FROM DATABASE country
JOIN DATABASE gender USING (uid)
JOIN DATABASE age USING (uid)
WHERE
country.fid = 3 AND
gender.fid = 6 AND
age.fid = 2 AND
county.value = 'spain' AND
gender.value = 'male' AND
age.value >= 20
age.value <= 30
GROUP BY uid
HAVING COUNT(*) >= 2
This code should work more reliably.
Use a self join:
SELECT tland.uid
FROM `table` AS tland
INNER JOIN `table` AS tgender ON tland.uid = tgender.uid
INNER JOIN `table` AS tage ON tland.uid = tage.uid
WHERE tland.value = 'spain'
AND tgender.value = 'male'
AND 20 <= tage.value AND tage.value <= 30
This will work.
SELECT d1.uid FROM demo d1
INNER JOIN demo d2 ON d2.uid = d1.uid AND d2.fid = 6 AND d2.value = "Male"
INNER JOIN demo d3 ON d3.uid = d1.uid AND d3.fid = 3 AND d3.value = "Spain"
WHERE (d1.fid = 2 AND (d1.value BETWEEN 20 AND 30))
Keeping your syntax as close to what it was as possible, this might work:
SELECT uid
FROM database
WHERE value in ('spain', 'male') or convert(value, , SIGNED INTEGER) between 30 and 50
GROUP BY uid HAVING COUNT(*) >= 3
Since your fields are mapped as so:
fid 3 = country
fid 2 = age
fid 6 = sex
You can transform that table into a more logical view (indexed on UID)
(SELECT c.uid, c.country, a.age, s.sex FROM
(SELECT uid, value AS country FROM maintable WHERE fid = 3) c INNER JOIN
(SELECT uid, value AS age FROM maintable WHERE fid = 2) a ON c.uid = a.uid INNER JOIN
(SELECT uid, value AS sex FROM maintable WHERE fid = 6) s ON c.uid = s.uid
) as X
Once it is inside a view you can then treat your data as a regular table. I'm assuming you may have more than 3 types of data since fid goes up to 6 on your example.
Assuming you have put that into a view (called X), you can just do:
SELECT * FROM X
WHERE age BETWEEN 20 AND 30
AND country = 'spain'
AND sex = 'male'