根据表中每个组的唯一列选择行

Question

 
1,     1,     1 - not elegible to select as next entry has same p_id
1,     2,     1
1,     3,     3

2,     1,     1
2,     2,     2
2,     3,     3

3,     1,     3 - not elegible to select as next entry has same p_id
3,     2,     3 - not elegible to select as next entry has same p_id
3,     3,     3

What is the cost effective way of writing the query to produce below result from above data ?

 
  1,    2,   1
  1,    3,   3

  2,    1,   1
  2,    2,   2
  2,    3,   3

  3,    3,   3

Answer 1

In MySQL 8.0, you can use lead() to retrieve the "next" p_id , and use that information to filter out rows whose pid is the same as the next value.

select *
from (
    select t.*, lead(p_id) over(partition by c_id order by s_id) lead_p_id
    from mytable t
) t
where not p_id <=> lead_p_id

In earlier versions, you would typically use a correlated subquery:

select t.*
from mytable t
where not pid <=> (
    select p_id
    from mytable t1
    where t1.c_id = t.c_id and t1.s_id > t.s_id
    order by t1.s_id
    limit 1
)

Answer 2

I'm not too sure if this is the most cost effective, but it def seems to be the most obvious solution I could think of.

select 
   c_id
   ,max(s_id) [s_id]
   ,p_id
from
  `table_name`
group by
   c_id
   ,p_id

Answer 3

If s_id is a sequence with no gaps, you can use:

select t.*
from t left join
     t tnext
     on tnext.c_id = t.c_id and tnext.s_id = t.sid + 1
where not (tnext.p_id <=> t.p_id);