Is there a better way to write this Oracle SQL that deals with comparing dates in two different columns?

Question

For a table in an Oracle 11g database:

CREATE TABLE mytable (START_DATE_TIME DATE, END_DATE_TIME DATE);
Insert into MYTABLE (START_DATE_TIME,END_DATE_TIME) values (to_date('06/20/2013 01:30:00 PM','MM/DD/YYYY HH:MI:SS AM'),to_date('06/20/2013 06:15:00 PM','MM/DD/YYYY HH:MI:SS AM'));
Insert into MYTABLE (START_DATE_TIME,END_DATE_TIME) values (to_date('06/21/2013 06:45:00 PM','MM/DD/YYYY HH:MI:SS AM'),to_date('06/22/2013 06:45:00 AM','MM/DD/YYYY HH:MI:SS AM'));
Insert into MYTABLE (START_DATE_TIME,END_DATE_TIME) values (to_date('06/23/2013 06:45:00 AM','MM/DD/YYYY HH:MI:SS AM'),to_date('06/23/2013 10:30:00 PM','MM/DD/YYYY HH:MI:SS AM'));
Insert into MYTABLE (START_DATE_TIME,END_DATE_TIME) values (to_date('06/25/2013 08:00:00 AM','MM/DD/YYYY HH:MI:SS AM'),to_date('06/25/2013 05:30:00 PM','MM/DD/YYYY HH:MI:SS AM'));
Insert into MYTABLE (START_DATE_TIME,END_DATE_TIME) values (to_date('06/25/2013 05:30:00 PM','MM/DD/YYYY HH:MI:SS AM'),to_date('06/26/2013 05:30:00 AM','MM/DD/YYYY HH:MI:SS AM'));

START_DATE_TIME         END_DATE_TIME
06/20/2013 01:30:00 PM  06/20/2013 06:15:00 PM
06/21/2013 06:45:00 PM  06/22/2013 06:45:00 AM
06/23/2013 06:45:00 AM  06/23/2013 10:30:00 PM
06/25/2013 08:00:00 AM  06/25/2013 05:30:00 PM
06/25/2013 05:30:00 PM  06/26/2013 05:30:00 AM

My goal is to find the hours between END_DATE_TIME and the next most recent START_DATE_TIME.

My query is :

SELECT start_date_time,
  end_date_time,
  ROUND((next_start_date - end_date_time)*24,2) hours_to_next_start
FROM
  (SELECT t.start_date_time,
    t.end_date_time,
    (SELECT MIN(t2.start_date_time)
    FROM mytable t2
    WHERE t2.start_date_time >= t.end_date_time
    ) next_start_date
  FROM mytable t
  ORDER BY t.start_date_time
  )

and I get the results:

START_DATE_TIME         END_DATE_TIME           HOURS_TO_NEXT_START 
06/20/2013 01:30:00 PM  06/20/2013 06:15:00 PM  24.5
06/21/2013 06:45:00 PM  06/22/2013 06:45:00 AM  24
06/23/2013 06:45:00 AM  06/23/2013 10:30:00 PM  33.5
06/25/2013 08:00:00 AM  06/25/2013 05:30:00 PM  0
06/25/2013 05:30:00 PM  06/26/2013 05:30:00 AM  NULL

It works fine as far as I know. But something tells me there might be a better, or at least newer, way to do this without the secondary query in the select?

What do you say?

Answer 1

There is a more "modern" way to do this, using the analytic LEAD function. I think it's better because it's more compact, plus once you get used to the analytic functions you'll find them easy to read - but that's just an opinion :)

SELECT
  Start_Date_Time,
  End_Date_Time,
  ROUND((next_start_date - end_date_time)*24,2) hours_to_next_start
FROM (
  SELECT
    Start_Date_Time,
    End_Date_Time,
    LEAD(Start_Date_Time) OVER (ORDER BY Start_Date_Time) AS   Next_Start_Date
  FROM myTable
)
ORDER BY Start_Date_Time

There's a SQL Fiddle here . To get more insight into how this works, try running the inner query on its own.

Answer 2

Use this query for getting diff hours between two date this will return as HH.MM

SELECT 
CONCAT
(
 DATEDIFF(MI,'06/20/2013 01:30:00 PM','06/20/2013 06:15:00 PM') / 60 ,'.', DATEDIFF(MI,'06/20/2013 01:30:00 PM','06/20/2013 06:15:00 PM') % 60
)