I have 2 tables in my database. One is for orders, and one is for companies.
Orders has this structure:
OrderID | attachedCompanyIDs
-------
because the second query is looking for rows with the id's 1 OR 2 OR 3, the first query is looking for a one of the comma delimited values to exist in companyID,
and another problem here is you aren't joining the tables on a common key in your where so you are going to get a mutation of rows that = count(table1) * count(table2);
Your problem really exists with part 2 of my answer. (with your second query)
SELECT o.*, GROUP_CONCAT(c.name) FROM Orders AS o , Company.c
WHERE FIND_IN_SET(c.CompanyID , o.attachedCompanyIDs) GROUP BY o.attachedCompanyIDs
attachedCompanyIDs is one big string, so mysql try to find company in this its cast to integer
when you use where in
so if comapnyid = 1 :
companyID IN ('1,2,3')
this is return true
but if the number 1 is not in the first place
companyID IN ('2,3,1')
its return false
To get the all related companies name, not based on particular Id.
SELECT
(SELECT GROUP_CONCAT(cmp.cmpny_name)
FROM company cmp
WHERE FIND_IN_SET(cmp.CompanyID, odr.attachedCompanyIDs)
) AS COMPANIES
FROM orders odr
SELECT name
FROM orders,company
WHERE orderID = 1
AND companyID IN (attachedCompanyIDs)
attachedCompanyIDs
is a scalar value which is cast into INT
(type of companyID
).
The cast only returns numbers up to the first non-digit (a comma in your case).
Thus,
companyID IN ('1,2,3') ≡ companyID IN (CAST('1,2,3' AS INT)) ≡ companyID IN (1)
In PostgreSQL
, you could cast the string into array (or store it as an array in the first place):
SELECT name
FROM orders
JOIN company
ON companyID = ANY (('{' | attachedCompanyIDs | '}')::INT[])
WHERE orderID = 1
and this would even use an index on companyID
.
Unfortunately, this does not work in MySQL
since the latter does not support arrays.
You may find this article interesting (see #2
):
Update:
If there is some reasonable limit on the number of values in the comma separated lists (say, no more than 5
), so you can try to use this query:
SELECT name
FROM orders
CROSS JOIN
(
SELECT 1 AS pos
UNION ALL
SELECT 2 AS pos
UNION ALL
SELECT 3 AS pos
UNION ALL
SELECT 4 AS pos
UNION ALL
SELECT 5 AS pos
) q
JOIN company
ON companyID = CAST(NULLIF(SUBSTRING_INDEX(attachedCompanyIDs, ',', -pos), SUBSTRING_INDEX(attachedCompanyIDs, ',', 1 - pos)) AS UNSIGNED)
Let me explain when to use FIND_IN_SET and When to use IN.
Let's take table A which has columns named "aid","aname". Let's take table B which has columns named "bid","bname","aids".
Now there are dummy values in Table A and Table B as below.
aid aname
1 Apple
2 Banana
3 Mango
bid bname aids
1 Apple 1,2
2 Banana 2,1
3 Mango 3,1,2
enter code here
Case1: if you want to get those records from table b which has 1 value present in aids columns then you have to use FIND_IN_SET.
Query: select * from A JOIN B ON FIND_IN_SET(A.aid,b.aids) where A.aid = 1 ;
Case2: if you want to get those records from table a which has 1 OR 2 OR 3 value present in aid columns then you have to use IN.
Query: select * from A JOIN B ON A.aid IN (b.aids);
Now here upto you that what you needs through mysql query.