Set a list in a variable in subquery - MYSQL - mysql

Question

My problem is the following, I want set a list of ID in a variable, then use this variable in a subquery. The problem is that WorkBench (my GUI) return the following error : "subquery returning multiple rows". It seems to me that's what I want.
Please explain me where I am wrong.
This is my query :
set #listID := (select ID_VOIE as ID from voies
where ORIGINE = 'XXX'
group by CODE_INSEE, CODE_VOIE
having count(*) > 1);
select substring(v.CODE_INSEE,1,2), count(*) from voies v
where v.ID_VOIE in (#listID)
group by substring(vs.CODE_INSEE,1,2);
The thing is I'm blocked with the "group by", I want do a groupd by after a first group by, that's why I can't (or at least i didn't find a way) write the request with a single WHERE clause.
The thing is I know that I can put the whole request directly in my subquery instead of using variable but :
It can let me use this trick in another requests that needed this behaviour (DRY concept !)
I'm not sure but the subquery will be executed in each turn of my loop, and that will be very inefficient
So I seek 2 possible ways : a way that let me use a list in a variable in a subquery OR a way that let me use "group by" twice in a single query.
Thanks you in advance for your answers (oh and sorry for my english, this is not my maternal language).

Answer 1

Unless you need that variable for something else, you should be able to skip it entirely as follows:
SELECT
SUBSTRING(v.CODE_INSEE,1,2),
COUNT(*)
FROM
voies v
WHERE
v.ID_VOIE in
(SELECT
ID_VOIE as ID
FROM
voies
WHERE
ORIGINE = 'XXX'
GROUP BY
CODE_INSEE,
CODE_VOIE
HAVING COUNT(*) > 1)
GROUP BY
SUBSTRING(vs.CODE_INSEE,1,2);
As you say, the subquery will be executed for all rows. To avoid that, a variable would be best, but MySQL doesn't support table variables. Instead, you can use a temporary table:
IF EXISTS DROP TABLE myTempTable;
CREATE TEMPORARY TABLE myTempTable (ID_VOIE int); -- I don't know the datatype
INSERT INTO myTempTable (ID_VOIE)
SELECT DISTINCT -- using distinct so I can join instead of use IN.
ID_VOIE as ID from voies
WHERE
ORIGINE = 'XXX'
GROUP BY
CODE_INSEE, CODE_VOIE
HAVING COUNT(*) > 1
And now you can do this:
SELECT
SUBSTRING(v.CODE_INSEE,1,2), COUNT(*)
FROM
voies v
JOIN myTempTable tt ON
v.ID_VOIE = tt.ID_VOIE
GROUP BY SUBSTRING(vs.CODE_INSEE,1,2);