Question

So I have a relatively complicated MySQL query that is giving a Duplicate Column, for reasons I haven’t been able to determine. I’m somewhat new to SQL.

The query is supposed to join a list of “problem” entries with the database of “tickets,” specifically the latest ticket for the particular id. Thus I have an inner query that executes a JOIN of the ticket table on itself to get the latest ticket. This works correctly. I then attempt to JOIN this inner selection with the problem database, however, I get a Duplicate Column Name id, despite the fact that this is the column name I am joining on. My understanding is that when you join on a particular column, it shouldn’t “count” as duplicate…

Anyway, the query:

SELECT * FROM `problem` P
LEFT JOIN (
    SELECT * FROM (
        SELECT id, MAX(ticket_id) AS ticket_id
        FROM `ticket`
        WHERE ticket_status = 'o'
        GROUP BY id
    ) AS MAX
    INNER JOIN (
        SELECT id, ticket_id, other_data, time_stamp FROM `ticket`
    ) CUR ON MAX.id = CUR.id AND MAX.ticket_id = CUR.ticket_id
) T ON P.id=T.id
WHERE
        P.since IS NOT NULL;

As I said, if I do just this:

SELECT * FROM (
    SELECT id, MAX(ticket_id) AS ticket_id
    FROM `ticket`
    WHERE ticket_status = 'o'
    GROUP BY id
) AS MAX
INNER JOIN (
    SELECT id, ticket_id, other_data, time_stamp FROM `ticket`
) CUR ON MAX.id = CUR.id AND MAX.ticket_id = CUR.ticket_id

I get a table of the latest ticket for each id.

One thing I have tried to fix this is to SELECT MAX.* FROM ( instead of just SELECT * FROM (. This eliminates the duplicate column name error, but the inner query no longer returns any other_data or time_stamp – I only get the id and ticket_id columns. This defeats the purpose of the join in the first place, of course.

Answer 1

Alias the id column as below:

SELECT * FROM `problem` P
LEFT JOIN (
    SELECT * FROM (
        SELECT id AS max_id, MAX(ticket_id) AS ticket_id
        FROM `ticket`
        WHERE ticket_status = 'o'
        GROUP BY max_id
    ) AS MAX
    INNER JOIN (
        SELECT id AS cur_id, ticket_id, other_data, time_stamp FROM `ticket`
    ) CUR ON MAX.max_id = CUR.cur_id AND MAX.ticket_id = CUR.ticket_id
) T ON P.id=T.max_id
WHERE
        P.since IS NOT NULL;

I wrote in a comment that you could just select the needed columns to avoid the duplicate and the question author understood I meant the external select, I was talking about the inner one. Hope I wrote it right, didn’t test.

SELECT * FROM `problem` P
LEFT JOIN (
    SELECT CUR.id, CUR.ticket_id, CUR.other_data, CUR.time_stamp FROM (
        SELECT id, MAX(ticket_id)
        FROM `ticket`
        WHERE ticket_status = 'o'
        GROUP BY max_id
    ) AS MAX
    INNER JOIN (
        SELECT id, ticket_id, other_data, time_stamp FROM `ticket`
    ) CUR ON MAX.max_id = CUR.cur_id AND MAX.ticket_id = CUR.ticket_id
) T ON P.id=T.id
WHERE
        P.since IS NOT NULL;