Do you wonder if your IT business is truly profitable or if you should raise your prices? Learn how to calculate your overhead burden using our free interactive tool and use it to determine the right price for your IT services. Start calculating Now!
SQL Server 2008: Query logic using CTE
Hello,
I've a staging table in which I've many rows with respect to booking number.
So, can I do in a single query using CTE to find out below concern? FYI, one way I can use cursor or some iterative loop. But I would like to use in single query using CTE.
--> as per below tables records, booking # 101, A needs to pay $3000 to B; So, first A paid to E = 1000 and A paid to C = 2000 (so that E & C now responsible to pay B).
So, E, C, D, F, X, Y, and Z all our mediator.
Below thing I need to achieve from this table:
1. How many mediator used in this booking and the amount they received from source point; actually they paid all amount to other mediator like A to E 1000 so, from E, it should be paid to other like D & F 500. It is same for C, X, Y & Z?
2. The amount should be delivered completely for each mediator i.e. it should not D = 500 and F = 490.
For e.g., (also, please find attached .png file here)
Note: Booking #, Source & Destination will be duplicate.
Booking#	Source	Dest
101	A	B	A	E	
101	A	B	A	C	
101	A	B	E	D	
101	A	B	E	F	
101	A	B	C	X	
101	A	B	C	Y	
101	A	B	C	Z	
101	A	B	X	B	
101	A	B	Y	B	
101	A	B	Z	B	
Best Regards,
Mohit Pandit
001-Snap.PNG
I've a staging table in which I've many rows with respect to booking number.
So, can I do in a single query using CTE to find out below concern? FYI, one way I can use cursor or some iterative loop. But I would like to use in single query using CTE.
--> as per below tables records, booking # 101, A needs to pay $3000 to B; So, first A paid to E = 1000 and A paid to C = 2000 (so that E & C now responsible to pay B).
So, E, C, D, F, X, Y, and Z all our mediator.
Below thing I need to achieve from this table:
1. How many mediator used in this booking and the amount they received from source point; actually they paid all amount to other mediator like A to E 1000 so, from E, it should be paid to other like D & F 500. It is same for C, X, Y & Z?
2. The amount should be delivered completely for each mediator i.e. it should not D = 500 and F = 490.
For e.g., (also, please find attached .png file here)
Note: Booking #, Source & Destination will be duplicate.
Booking#	Source	Dest
101	A	B	A	E	
101	A	B	A	C	
101	A	B	E	D	
101	A	B	E	F	
101	A	B	C	X	
101	A	B	C	Y	
101	A	B	C	Z	
101	A	B	X	B	
101	A	B	Y	B	
101	A	B	Z	B	
Best Regards,
Mohit Pandit
001-Snap.PNG
Do more with
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Can you post here what output you required. ?
I mean a sample output dataset as per above data.... and also some more explanation....
I mean a sample output dataset as per above data.... and also some more explanation....
I need to put validations either respective booking is valid or not i.e. A needs to pay $3000 to B and the amount mediator receives then the same amount should travel to other mediator and at last the destination i.e. B.
Let's take A (source) paid E (mediator) 1000 then E paid to D & F (mediator) 500 each.
--> in that case E receives 1000 and same amount he paid to E & D i.e. 500 each.
I apologize, two more records should be in table; added in below table in last (marked with *).
Booking# Source Destination Mid Source Mid Destination Payment ($)
101 A B A E 1000
101 A B A C 2000
101 A B E D 500
101 A B E F 500
101 A B C X 1000
101 A B C Y 700
101 A B C Z 300
101 A B X B 1000
101 A B Y B 700
101 A B Z B 300
101 A B D B 500 *
101 A B F B 500 *
Similar A paid to C and C paid to X,Y, & Z and X,Y and Z paid to B. So, I need to put validations the amount given by source/mid source then it should be same travel at destination/mid destination otherwise I need to put an alert to end user.
I hope, it should clear now. Kindly let me know in case still not clear.
Let's take A (source) paid E (mediator) 1000 then E paid to D & F (mediator) 500 each.
--> in that case E receives 1000 and same amount he paid to E & D i.e. 500 each.
I apologize, two more records should be in table; added in below table in last (marked with *).
Booking# Source Destination Mid Source Mid Destination Payment ($)
101 A B A E 1000
101 A B A C 2000
101 A B E D 500
101 A B E F 500
101 A B C X 1000
101 A B C Y 700
101 A B C Z 300
101 A B X B 1000
101 A B Y B 700
101 A B Z B 300
101 A B D B 500 *
101 A B F B 500 *
Similar A paid to C and C paid to X,Y, & Z and X,Y and Z paid to B. So, I need to put validations the amount given by source/mid source then it should be same travel at destination/mid destination otherwise I need to put an alert to end user.
I hope, it should clear now. Kindly let me know in case still not clear.
Check below code... below query can solve your problem.. you can write it in your way...
even you can use CTE if you are working with multiple sources and destinations...
even you can use CTE if you are working with multiple sources and destinations...
Declare @ta table (Booking int, Source char(2) , Destination char(2) , [Mid Source] char(2) , [Mid Destination] char(2) , Payment int)
insert into @ta
select 101 , 'A' ,'B', 'A' , 'E' ,1000
union all select 101 , 'A' ,'B', 'A' ,'C' , 2000
union all select 101 , 'A' ,'B', 'E' ,'D' , 500
union all select 101 , 'A' ,'B', 'E' ,'F' , 500
union all select 101 , 'A' ,'B', 'C' ,'X' , 1000
union all select 101 , 'A' ,'B', 'C' ,'Y' , 700
union all select 101 , 'A' ,'B', 'C' ,'Z', 300
union all select 101 , 'A' ,'B', 'X' ,'B', 1000
union all select 101 , 'A' ,'B', 'Y' ,'B', 700
union all select 101 , 'A' ,'B', 'Z' ,'B', 300
union all select 101 , 'A' ,'B', 'D' ,'B', 500
union all select 101 , 'A' ,'B', 'F' ,'B', 50
select * from @ta
select t.*, case when x.payment<>t.Payment then 'Payment Mistmatch ' else '' end as [status] from @ta t join
(select booking,Source,Destination,[Mid Source] , SUM(Payment) as payment from @ta
group by booking,Source,Destination,[Mid Source]) x
on t.Booking=x.booking and t.Source=x.Source and t.Destination=x.Destination and t.[Mid Destination]=x.[Mid Source]
Sauv,
I had prepared something similar. It might suffice, but for one point - if a transaction is completely missing, it does not appear. E.g. if E made no transfer, A -> E with 1000 will not appear. Using a Left Join (and applying isnull to x.payment) should help.
Additionally I would also output the missing amount.
I had prepared something similar. It might suffice, but for one point - if a transaction is completely missing, it does not appear. E.g. if E made no transfer, A -> E with 1000 will not appear. Using a Left Join (and applying isnull to x.payment) should help.
Additionally I would also output the missing amount.
Hi Saurv,
I checked the code and looks good. Actually, I can have 30k records in a file so it would good if we will use CTE. Could you please share your thoughts?
I also noted Qlemo point for using LEFT OUTER JOIN.
Thanks
Best Regards,
Mohit Pandit
I checked the code and looks good. Actually, I can have 30k records in a file so it would good if we will use CTE. Could you please share your thoughts?
I also noted Qlemo point for using LEFT OUTER JOIN.
Thanks
Best Regards,
Mohit Pandit
Your original request is still not clear. If you have 30k records, what do you expect as output from them? Are interested in just in something like
Booking Source Dest Total Remaining NoMediators
101 A B 3000 0 7
Since you have 30K records then better would be to use a temp table to store a grouped sum records..
like this...
Declare @ta table (Booking int, Source char(2) , Destination char(2) , [Mid Source] char(2) , [Mid Destination] char(2) , Payment int)
Create table #temp_group
(
Booking int,
Source char(2),
Destination char(2),
[Mid Source] char(2) ,
payement_sum bigint
)
insert into @ta
select 101 , 'A' ,'B', 'A' , 'E' ,1000
union all select 101 , 'A' ,'B', 'A' ,'C' , 2000
union all select 101 , 'A' ,'B', 'E' ,'D' , 500
union all select 101 , 'A' ,'B', 'E' ,'F' , 500
union all select 101 , 'A' ,'B', 'C' ,'X' , 1000
union all select 101 , 'A' ,'B', 'C' ,'Y' , 700
union all select 101 , 'A' ,'B', 'C' ,'Z', 300
union all select 101 , 'A' ,'B', 'X' ,'B', 1000
union all select 101 , 'A' ,'B', 'Y' ,'B', 700
union all select 101 , 'A' ,'B', 'Z' ,'B', 300
union all select 101 , 'A' ,'B', 'D' ,'B', 500
union all select 101 , 'A' ,'B', 'F' ,'B', 50
union all select 101 , 'A1' ,'B1', 'A1' , 'E1' ,2000
union all select 101 , 'A1' ,'B1', 'A1' ,'C1' , 4000
union all select 101 , 'A1' ,'B1', 'E1' ,'D1' , 1000
union all select 101 , 'A1' ,'B1', 'E1' ,'F1' , 1000
union all select 101 , 'A1' ,'B1', 'C1' ,'X1' , 2000
union all select 101 , 'A1' ,'B1', 'C1' ,'Y1' , 1400
union all select 101 , 'A1' ,'B1', 'C1' ,'Z1', 600
union all select 101 , 'A1' ,'B1', 'X1' ,'B1', 2000
union all select 101 , 'A1' ,'B1', 'Y1' ,'B1', 1400
union all select 101 , 'A1' ,'B1', 'Z1' ,'B1', 600
union all select 101 , 'A1' ,'B1', 'D1' ,'B1', 1000
union all select 101 , 'A1' ,'B1', 'F1' ,'B1', 00
insert into #temp_group
select booking,Source,Destination,[Mid Source] , SUM(Payment) as payment from @ta
group by booking,Source,Destination,[Mid Source]
select t.*, case when x.payement_sum<>t.Payment then 'Payment Mistmatch ' else '' end as [status] from @ta t left join #temp_group x
on t.Booking=x.booking and t.Source=x.Source and t.Destination=x.Destination and t.[Mid Destination]=x.[Mid Source]
drop table #temp_group
like this...
select t.* from @ta t left join #temp_group x
on t.Booking=x.booking and t.Source=x.Source and t.Destination=x.Destination and t.[Mid Destination]=x.[Mid Source]
where x.payement_sum<>t.Payment
Ummm... with the left join there can be NULL where there is no payment from a mediator (ie they have received, but not handed on) and the code is not checking the NULL condition...
And, it is only going 1 level deep, it isnt going all the way as if it were a multi-level BOM (hierarchical data). For example, if we are joining on t.[mid destination] to x.[mid source] then why would we expect the t.[mid destination] of 'B' ever match to anything from the x.[mid source] ? So, (if we handle NULL in that case statement), it would report as a mismatch.
Given the original example (with the missing D and F payments to B) then we should see that B has been short paid by 1000 and D and F have both received 500 with no subsequent payments.
qlemo does raise a couple of interesting questions (as well as noting the left join condition)...
If we add in the two additional items (those marked with *) then it all balances just fine, so, what do we want to see ?
We almost have to look at all payments and receipts by individual mediator, and if they pay everything they receive, then we dont have to worry about that mediator.
So, lets look at what happens with NULL handling and then, another approach (using CTE) depending on what you actually want to see :
And, we could probably go down the BOM / hierarchy type approach where you get to see the path of transactions and where the breakdown was, or, some kind of summary level...
And, it is only going 1 level deep, it isnt going all the way as if it were a multi-level BOM (hierarchical data). For example, if we are joining on t.[mid destination] to x.[mid source] then why would we expect the t.[mid destination] of 'B' ever match to anything from the x.[mid source] ? So, (if we handle NULL in that case statement), it would report as a mismatch.
Given the original example (with the missing D and F payments to B) then we should see that B has been short paid by 1000 and D and F have both received 500 with no subsequent payments.
qlemo does raise a couple of interesting questions (as well as noting the left join condition)...
If we add in the two additional items (those marked with *) then it all balances just fine, so, what do we want to see ?
We almost have to look at all payments and receipts by individual mediator, and if they pay everything they receive, then we dont have to worry about that mediator.
So, lets look at what happens with NULL handling and then, another approach (using CTE) depending on what you actually want to see :
Declare @ta table (Booking int, Source char(2) , Destination char(2) , [Mid Source] char(2) , [Mid Destination] char(2) , Payment int)
insert into @ta
select 101 , 'A' ,'B', 'A' , 'E' ,1000
union all select 101 , 'A' ,'B', 'A' ,'C' , 2000
union all select 101 , 'A' ,'B', 'E' ,'D' , 500
union all select 101 , 'A' ,'B', 'E' ,'F' , 500
union all select 101 , 'A' ,'B', 'C' ,'X' , 1000
union all select 101 , 'A' ,'B', 'C' ,'Y' , 700
union all select 101 , 'A' ,'B', 'C' ,'Z', 300
union all select 101 , 'A' ,'B', 'X' ,'B', 1000
union all select 101 , 'A' ,'B', 'Y' ,'B', 700
union all select 101 , 'A' ,'B', 'Z' ,'B', 300
--union all select 101 , 'A' ,'B', 'D' ,'B', 500 --*
--union all select 101 , 'A' ,'B', 'F' ,'B', 500 --*
-- let's put in a bit of NULL handling in the CASE statement
select t.*, case when x.payment is NULL or x.payment<>t.Payment then 'Payment Mistmatch ' else '' end as [status] , x.payment as source_payments
from @ta t
left join (select booking,Source,Destination,[Mid Source] , SUM(Payment) as payment from @ta group by booking,Source,Destination,[Mid Source]) x
on t.Booking=x.booking and t.Source=x.Source and t.Destination=x.Destination and t.[Mid Destination]=x.[Mid Source]
-- now lets try to match up on payments and receipts by mediator
;with cte_rec as
(
select booking, source, destination, case when [mid source] = source then '' else [mid source] end as mediator, payment * -1 as payment
from @ta
union all
select booking, source, destination, case when [mid destination] = destination then '' else [mid destination] end, payment
from @ta
)
select booking, source, destination, mediator, sum(payment) as balance
from cte_rec
-- where mediator > '' -- taken out to see total still owed to destination
group by booking, source, destination, mediator
having sum(payment) <> 0
And, we could probably go down the BOM / hierarchy type approach where you get to see the path of transactions and where the breakdown was, or, some kind of summary level...
Premium Content
You need an Expert Office subscription to comment.Start Free Trial