troubleshooting Question

how to do multiple partitioning

Avatar of lisa_mc
lisa_mcFlag for United Kingdom of Great Britain and Northern Ireland asked on
DatabasesMicrosoft SQL Server 2005SQL
6 Comments1 Solution211 ViewsLast Modified:
Hi everyone

I was wondering if its possible to partition a set of data then get this data and partition it again?

My scenarios are
1) Need to firstly check that a register which has the same date, time, staff_code, week_no, and course I want to only pick the modules beginning with m (b.aos_code sorted desc)
2) Having carried out the first check I then need check the data for registers that occur same date, time, staff_code, week_no which have two courses attached in this case I want to select the highest nvq_lvl and if the nvq_lvls are the same then check the hrs_total

I have seperated both scenarios into partitions but im not sure if you can do what I have attached below ie do you need to partition and then store in a table and then partition again?
(I have removed the tables to shorten the code)

Help appreciated
select acad_period, register_id, register_group, staff_code, week_no, day_of_week, date, [month_no],
start_time, end_time,
convert(varchar, convert(int,floor([Total Hours])))+':'+
replicate('0',(2 - len(convert (varchar, convert(int,([Total Hours] - floor([Total Hours])) * 60.0)))))
+convert (varchar, convert(int,([Total Hours] - floor([Total Hours])) * 60.0)) as 'Total Hours',
dept_code, 
[m_code], [m_period], [m_desc], [c_code], [c_period], [c_desc], 
[c_moa], [c_hrs], nvq_lvl_ind, [rn], [rn1]


FROM
(
			select a.acad_period, a.register_id, a.register_group, a.staff_code,
			a.week_no,
			c.day_of_week, g.date,
			left(CONVERT ( char(20) , g.date , 101 ),2) as 'month_no',
			c.start_time, c.end_time,
			ROUND(cast((datediff(mi, MIN(c.start_time), MAX(c.end_time))  / 60.0) as FLOAT),2) AS 'Total Hours',
			c.dept_code, 
			b.aos_code as 'm_code', b.aos_period as 'm_period', f.full_desc as 'm_desc',
			e.aos_code as 'c_code', e.aos_period as 'c_period', e.full_desc as 'c_desc', e.moa_code as 'c_moa',
			e.hrs_total as 'c_hrs',
			h.nvq_lvl_ind,
			row_number() over (partition by a.register_id, a.register_group, a.staff_code, a.week_no, c.day_of_week, g.date,
							c.start_time, c.end_time, e.aos_code, e.aos_period
                         order by b.aos_code desc ) as rn ,
			row_number() over (partition by a.register_id, a.register_group, a.staff_code, a.week_no, c.day_of_week, g.date,
							c.start_time, c.end_time
                         order by h.nvq_lvl_ind, e.hrs_total desc ) as rn1 

			from tables ....

			where a.acad_period = '10/11'

			group by a.acad_period, a.register_id, a.register_group, a.staff_code, a.week_no,
			c.day_of_week, g.date, c.start_time,
			c.end_time, c.dept_code, 
			b.aos_code, b.aos_period, f.full_desc,
			e.aos_code, e.aos_period, e.full_desc, 
			e.moa_code, e.hrs_total, h.nvq_lvl_ind

) X

where (rn = 1 and rn1 = 1)

order by staff_code, [c_code], [c_period], register_id, register_group, date
Join the community to see this answer!
Join our exclusive community to see this answer & millions of others.
Unlock 1 Answer and 6 Comments.
Join the Community
Learn from the best

Network and collaborate with thousands of CTOs, CISOs, and IT Pros rooting for you and your success.

Andrew Hancock - VMware vExpert
See if this solution works for you by signing up for a 7 day free trial.
Unlock 1 Answer and 6 Comments.
Try for 7 days

”The time we save is the biggest benefit of E-E to our team. What could take multiple guys 2 hours or more each to find is accessed in around 15 minutes on Experts Exchange.

-Mike Kapnisakis, Warner Bros