Question

i have a table with jobs like this in a MS SQL Server 2008 database:

J_ID      J_Name       J_ReadDate                  J_I_ID
`````````````````````````````````````````````````````````
01        Job1        2011-06-26 01:00:12.153      1
02        Job2        2011-06-25 01:00:07.103      1
03        Job3        2011-06-25 01:00:07.103      2
04        Job4        2011-06-25 01:00:07.103      3
05        Job1        2011-06-26 01:00:07.103      1
06        Job2        2011-06-26 01:00:08.103      2
07        Job1        2011-06-26 01:00:08.103      2

now i want to select only the newest jobs, no matter how old the entry is. So in my example it would like to get this result (ordered by J_ReadDate DESC):

J_ID      J_Name       J_ReadDate                  J_I_ID
`````````````````````````````````````````````````````````
06        Job2        2011-06-26 01:00:08.103      2
07        Job1        2011-06-26 01:00:08.103      2
05        Job1        2011-06-26 01:00:07.103      1
03        Job3        2011-06-25 01:00:07.103      2
04        Job4        2011-06-25 01:00:07.103      3

how can I achieve this? I cannot say, that the time difference between the entries is always one day.

Answer 1

The below will get what you’re after.

SELECT J_ID, J_Name, J_ReadDate, J_I_ID
FROM (
  SELECT J_ID, J_Name, J_ReadDate, J_I_ID,
         ROW_NUMBER() OVER (PARTITION BY J_Name ORDER BY J_ReadDate DESC) AS RN
  FROM jobs
) J
WHERE J.RN = 1
ORDER BY J_ReadDate DESC