使用IN与group_concat结果进行MySQL查询

Question

I'm trying to clean a db with duplicate records. I need to move the reference to a single record and delete the other one.

I have two tables: Promoters and Venues, each has a reference to a table called cities. The problem is that there are cities with the same name and different ids, that have a relation with venues and promoters.

With this query I can group all promoters and venues with a single city record:

SELECT c.id as id, c.name as name, GROUP_CONCAT( DISTINCT p.id ) as promoters_ids, GROUP_CONCAT( DISTINCT v.id ) as venues_ids
FROM cities as c
LEFT JOIN promoters as p ON p.city_id = c.id
LEFT JOIN venues as v ON v.city_id = c.id
WHERE c.name IN ( SELECT name from cities group by name having count(cities.name) > 1 )
GROUP BY c.name

Now I want to run an UPDATE query on promoters, setting the city_id equals to the result of the query above.

Something like this:

    UPDATE promoters AS pr SET pr.city_id = (
        SELECT ID
        FROM (
            SELECT c.id as id, c.name as name, GROUP_CONCAT( DISTINCT p.id ) as promoters_ids
            FROM cities as c
            LEFT JOIN promoters as p ON p.city_id = c.id

            WHERE c.name IN ( SELECT name from cities group by name having count(cities.name) > 1 ) AND pr.id IN promoters_ids
            GROUP BY c.name
            ) AS T1 

    )

How can I do this?

Thanks

Answer 1

If I understand correctly, you want to remove duplicate cities (in the end), so you need to update promoters that are linked to any of the cities you want to remove in that process.

I think it makes sense to use the lowest ID of any of the cities with the same name (could be the highest just as well, but I want to specify it at least, and don't leave it up to me.

So in order get the right ID for a promoter, I need to: Select the lowest ID of all cities that have the same name as the city already linked to a promoter.

Fortunately, that demand fits snuggly into a query:

UPDATE promoters AS pr 
SET pr.city_id = (
  SELECT 
    -- Select the lowest ID ..
    Min(c.id)
  FROM
    -- .. of all cities ..
    Cities c
    -- .. that have the same name ..
    INNER JOIN Cities pc on pc.Name = c.Name
  WHERE
    .. as the city already linked to the promoter being updated
    pc.id = pr.city_id
  GROUP BY
    c.name)

The trick is to join Cities on itself by name, so you can easily get all cities with the same name. I think you tried the same with the IN clause, but that's a little more complex than it needs to be.

I don't think you need group_concat at all, besides checking if the inned query returns the correct cities indeed, although it doesn't make sense, since you're already grouping on the name. When written like this, you can tell that there should be no way that this can go wrong:

  SELECT 
    -- Select the lowest ID ..
    MIN(c.id) AS id,
    GROUP_CONCAT(c.name) AS names --< already grouped by this, so why...
  FROM
    -- .. of all cities ..
    Cities c
    -- .. that have the same name.
    INNER JOIN Cities pc on pc.Name = c.Name
  GROUP BY
    c.name

I hope I understood the question correctly.