Home/ Questions/Q 8725753
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-06-13T08:06:16+00:00

I’m am trying to join a table that has a boolean column, and in

  • 0

I’m am trying to join a table that has a boolean column, and in my results I’m trying to get two columns from the join, one where the boolean is 0 and a separate on where it is 1.

Since there is a one to many relationship, I need to do a GROUP BY.

The following query returns the correct non GROUP BY results.

For each profile ID there is at least one row for both genre and influence.

SELECT p.profileID, IF( gpro.isInfluence =0, g.genreName, NULL ) AS genre, IF( gpro.isInfluence =1, g.genreName, NULL ) AS influence, g.genreName
FROM profiles p
LEFT JOIN genre_profiles gpro ON gpro.profileID = p.profileID
LEFT JOIN genres g ON g.genreID = gpro.genreID
WHERE g.genreName IS NOT NULL

However, when I do the corresponding GROUP BY/GROUP_CONCAT the output does not make sense.

SELECT p.profileID, IF( gpro.isInfluence =0, GROUP_CONCAT( g.genreName ) , NULL ) AS genre, IF( gpro.isInfluence =1, GROUP_CONCAT( g.genreName ) , NULL ) AS influence, g.genreName
FROM profiles p
LEFT OUTER JOIN genre_profiles gpro ON gpro.profileID = p.profileID
LEFT OUTER JOIN genres g ON g.genreID = gpro.genreID
WHERE g.genreName IS NOT NULL 
GROUP BY p.profileID

The above groups all the g.genreName rows in either of the columns, instead of them being split based on the gpro.isInfluence value. The column the result falls under I believe corresponds to the first result that is grouped.

1000001052  Latin American,Rock,Hip Hop NULL
1000001637  NULL    Electronic,Easy listening,Asian,popgun
1000001666  NULL    Electronic,Bacon,Emo,Modern folk,Hip Hop,R...

What I don’t understand is why are the genre values are grouped together and not in their respective columns like the first query.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team
    SELECT p.profileID, GROUP_CONCAT( IF( gpro.isInfluence =0, g.genreName, NULL ) ) AS genre, GROUP_CONCAT( IF( gpro.isInfluence =1, g.genreName, NULL ) ) AS influence, g.genreName
FROM profiles p
LEFT OUTER JOIN genre_profiles gpro ON gpro.profileID = p.profileID
LEFT OUTER JOIN genres g ON g.genreID = gpro.genreID
WHERE g.genreName IS NOT NULL 
GROUP BY p.profileID

    Putting the GROUP_CONCAT outside the IF statement gives the correct output, thanks to Ertunc for helping my brain make the pivot on that one.

    • 0