SQL query to return a specific row from a group

Question

Using SQL Server

The problem I have is best described by looking at the table below, I want to return all the rows highlighted red. I need to group all the rows by internal_id as this can be repeated several times and from that result set display the row with the smallest id value. There is a further restriction that the dataset must = 'TEST1' and the type = 'CHANGE'.

Answer 1

Use row_number() function :

select top (1) with ties t.*
from table t
where dataset = 'test1' and type = 'change'
order by row_number() over (partition by internal_id order by id);

Answer 2

added with test data.

create table #temp_1
    ( dataset varchar(7) null
    ,id int null
    ,duration int null
    ,type varchar(10) null
    ,internal_id int null
    )


    insert into #temp_1 values
    ('TEST1',97,61,'cOMPLETE',7)
    ,('TEST1',98,61,'cHANGE',13)
    ,('TEST1',102,61,'cHANGE',16)
    ,('TEST1',103,611,'cHANGE',15)
    ,('TEST1',107,601,'cHANGE',5)
    ,('TEST1',110,601,'cHANGE',25)
    ,('TEST1',111,613,'cHANGE',35)
    ,('TEST1',113,615,'cHANGE',45)
    ,('TEST1',193,619,'cHANGE',5)
    ,('TEST1',200,620,'cHANGE',51)

    select *
    from (
    SELECT *
    ,Rank_dense = dense_rank() over(partition by type order by internal_id asc)
    ,Rank_regular = rank() over(partition by type order by internal_id asc)
    from #temp_1
    where dataset = 'Test1' and type = 'change'
    ) a
    where Rank_dense = 1 -- or Rank_dense