SQL join to join values in primary table with child table values as different columns in a single row

Question

I have a survey_datas table contain data's as like this

survey_data_id  | title
1               | Paul 
3               | Anna 
4               | Alan

Another table project_playlist_indexes contain the data's like this

survey_id  |survey_data_id  | favorite
1          | 1              | 22.10
2          | 1              | 24.00
3          | 3              | 12.00

I want to join the survey_datas table with the project_playlist_indexes table so that values contained in the project_playlist_indexes table with same survey_data_id as survey_datas table should get as favorite time1 , favorite time 2, ... favorite time n , the result table I like to get is like this

survey_data_id  |title | favorite_time1 | favorite_time2
            1   | paul | 22.10          |24.00
            3   | anna | 12.00          | null
            4   | alan | null           | null

Currently I am using the query

SELECT s.*,GROUP_CONCAT(pi.favorite) ,pi.*
FROM survey_datas s
LEFT JOIN  project_playlist_indexes pi 
ON pi.survey_data_id = s.survey_data_id  
GROUP BY pi.survey_data_id

but the favorite values is getting in a single field I want it to be in different columns. How can I do this

score 1 · Accepted Answer · answered Nov 28 '17 at 11:54

You can do it by executing a dynamic sql query. What I have done is, first given a row number based on the survey_data_id column. Then selected each row number item as each column group by survey_data_id. Don't know how efficient the code is.

Query

set @query = null;
select
  group_concat(distinct
    concat(
      'max(case when `rn` = ',
      `rn`,
      ' then `favorite` end) as `favorite', `rn` , '`'
    )
  ) into @query
from (
  select `survey_id`, `survey_data_id`, `favorite`, (
    case `survey_data_id` when @curA 
    then @curRow := @curRow + 1 
    else @curRow := 1 and @curA := `survey_data_id` end 
  ) as `rn`
  from `project_playlist_indexes` t, 
  (select @curRow := 0, @curA := '') r 
  order by `survey_data_id`, `survey_data_id`
) t;

set @query = concat('select t2.`survey_data_id`, t2.`title`,', 
                @query,
              ' from (select `survey_id`, `survey_data_id`, `favorite`, (
              case `survey_data_id` when @curA 
              then @curRow := @curRow + 1 
              else @curRow := 1 and @curA := `survey_data_id` end 
              ) as `rn`
              from `project_playlist_indexes` t, 
              (select @curRow := 0, @curA := '''') r 
              order by `survey_data_id`, `survey_data_id`) t1
              right join `survey_datas` t2
              on t1.survey_data_id = t2.`survey_data_id`
              group by t1.`survey_data_id`
              order by t2.`survey_data_id`;'
     );

prepare stmt from @query;
execute stmt;
deallocate prepare stmt;

Find a demo here

SQL join to join values in primary table with child table values as different columns in a single row

1 Answers1