Question

I have a single table that looks like this:

Table: id, name

A name can show up many times. I’m trying to figure out an optimal query that will return for each name, its highest ID.

input data:

1, "a"
2, "a"
3, "b"
4, "a"
5", "b"

output:

4, "a"
5, "b"

I know of a way to do this using a subquery in the FROM portion of the query. I’m hoping I can somehow do faster.

I currently do:

SELECT table.* 
FROM table, (SELECT MAX(id) maxid FROM table GROUP BY name) maxids 
WHERE table.id = maxids.maxid

I am curious to see if there are faster ways!

Answer 1

Why use the self-join, this should work just fine:

select max(id), name
from yourtable
group by name

See SQL Fiddle with Demo

Now if you have more columns, that you want to include in your final result then you can use the following:

select t1.id, t1.name, t1.col1
from yourtable t1
inner join
(
  select max(id) id, name
  from yourtable
  group by name
) t2
  on t1.id = t2.id
  and t1.name = t1.name

Or even:

select t1.id, t1.name, t1.col1
from yourtable t1
where t1.id in (select max(t2.id)
                from yourtable t2
                group by t2.name);

See SQL Fiddle with Demo of both