Question

Given this exemplary relation and data

|_id_____|_type____|_name________|_date_______________|
| 1      | 1       | ajshdf      | 2012-02-18 12:35:00
| 2      | 1       | jkhae       | 2012-02-18 12:35:00
| 3      | 1       | dsfad       | 2012-02-18 12:35:00
| 4      | 2       | 3f3ad       | 2012-02-18 12:35:00
| 5      | 2       | gad64       | 2012-02-18 12:35:00
| 6      | 2       | as3da       | 2012-02-18 12:35:00
| 7      | 3       | faf3a       | 2012-02-18 12:35:00
| 8      | 3       | ga3d3       | 2012-02-18 12:35:00
| 9      | 4       | jzd64       | 2012-02-18 12:35:00

What is the best way to achieve a resulset that will distribute the records as evenly as possible, ordered by the types?

What I want is that the result records are ordered in a manner that the newest entries are always on top, but that the types are altered around, so that the result looks something like this:

|_id_____|_type____|_name________|_date_______________|
| 1      | 1       | ajshdf      | 2012-02-18 12:35:00
| 4      | 2       | 3f3ad       | 2012-02-18 12:35:00
| 7      | 3       | faf3a       | 2012-02-18 12:35:00
| 9      | 4       | jzd64       | 2012-02-18 12:35:00
| 2      | 1       | jkhae       | 2012-02-18 12:35:00
| 5      | 2       | gad64       | 2012-02-18 12:35:00
| 8      | 3       | ga3d3       | 2012-02-18 12:35:00
| 3      | 1       | dsfad       | 2012-02-18 12:35:00
| 6      | 2       | as3da       | 2012-02-18 12:35:00
| 7      | 3       | faf3a       | 2012-02-18 12:35:00 

As you see, the type is altered and cycled in the result – 1, 2, 3, 4, 1, 2, 3, 1, 2, 3

Answer 1

Pretty inefficient but I think it will work:

SELECT a.*
FROM 
       TableX AS a
   JOIN 
       TableX AS b
     ON  b.type = a.type
     AND b.id <= a.id 
GROUP BY 
      a.id
ORDER BY
      COUNT(*)
    , a.id