Updating multiple records from the previous row

What I want in the output is that if the closing balance then the op balance for the same date will become closing balance

having problems understanding this...

maybe this:

(case when closing is not null then closing else opening end) as closingBalance

just add this into your select...

What I want in the output is that if the there is no closing balance for the date then the op balance for the same date will become closing balance

i want a view for the same.

Hi,
It might be a better idea to have the backup of your original table before trying to run the code.
Assumptions i have taken:
1) The DivisionID, Opening & Closing are Integer
2) The LedgerId is same in all the records
3) The data is clustered index on LedgerId, DivisionId, ClosingDate (This will make sure the ordering is proper during the calculation)

Now try the below code:

declare @DivisionId int, @Opening int, @Closing int
select  @DivisionId = 0, @Opening = 0, @Closing = 0 
update myTable  
set 	Opening = case when isnull(opening,0) = 0 then @opening else opening end,
	Closing = case when isnull(closing,0) = 0 then @closing else closing end, 
	@opening = case when DivisionId = @DivisionId then @closing else 0 end, 
	@opening = case when isnull(opening,0) = 0 then @opening else opening end, 
	@closing = case when isnull(closing,0) = 0 then @opening else closing end,
	@Divisionid = divisionid 
select * from myTable

Select allOpen in new window

I think you got it wrong. The table data will remain the same. I would like a view for the data from the table as mentioned in the xl output section.

I am afraid, in a single direct SQL statement it is not possible as per my knowledge.
Yes you can create a multi-line table valued function. I have written the below code considering that you will call this function by passing the LedgerId.

create function dbo.retTable(@LedgerId Int) 
	returns @table table (LedgerId int, DivisionId int, ClosingDate date, opening int, closing int) 
as 
begin
	declare @DivisionId int, @Opening int, @Closing int
	select  @DivisionId = 0, @Opening = 0, @Closing = 0 
	insert into @table 
	select * from myTable where LedgerId = @LedgerId  
	update @table 
	set 	Opening = case when isnull(opening,0) = 0 then @opening else opening end,
		Closing = case when isnull(closing,0) = 0 then @closing else closing end, 
		@opening = case when DivisionId = @DivisionId then @closing else 0 end, 
		@opening = case when isnull(opening,0) = 0 then @opening else opening end, 
		@closing = case when isnull(closing,0) = 0 then @opening else closing end,
		@Divisionid = divisionid 
	return
end

Select allOpen in new window

And then call the function like this:

select * from dbo.retTable(101)

Select allOpen in new window

can you use a common table expression and make a view

I am not sure for the current case.

can somebody please help me on this its very urgent...

is this what you want?

with b as (
select 101 ledgerid, 1 divisionid, '1-Apr-11' Closingdate, 10001 opening, null closing
union all select 101,	1,	'2-Apr-11', null, null		
union all select 101,	1,	'3-Apr-11', 2500, null		
union all select 101,	1,	'4-Apr-11', null, 4250	
union all select 101,	1,	'5-Apr-11', null, null
union all select 101,	1,	'6-Apr-11',null,60000
)
select ledgerid, divisionid, Closingdate, opening, closing,
ISNULL(opening, pre_closing) opening2, 
isnull(ISNULL(closing, opening),pre_closing) closing2
from (
select 
*,
(select top 1 isnull(closing, opening) from b b2 where b1.ledgerid=b2.ledgerid and b1.divisionid=b2.divisionid and b2.closingdate<b1.closingdate and isnull(b2.closing, b2.opening) is not null order by Closingdate desc) pre_closing
from b b1
) x

ledgerid	divisionid	Closingdate	opening	closing	opening2	closing2
101	1	1-Apr-11	10001	NULL	10001	10001
101	1	2-Apr-11	NULL	NULL	10001	10001
101	1	3-Apr-11	2500	NULL	2500	2500
101	1	4-Apr-11	NULL	4250	2500	4250
101	1	5-Apr-11	NULL	NULL	4250	4250
101	1	6-Apr-11	NULL	60000	4250	60000

Select allOpen in new window

Good one HainCurt,
Just got reminded about the self joining and written below code to provide the result:

select x.ledgerid, x.divisionid, x.closingdate, isnull(x.opening,x.preclosing) new_opening, 
		isnull(isnull(x.closing,x.opening), preclosing) new_closing 
from (
select a.*, (select top 1 isnull(closing,opening) from t1 b 
		where a.ledgerid = b.ledgerid and a.divisionid = b.divisionid 
		and b.closingdate < a.closingdate and isnull(b.closing,b.opening) is not null 
		order by b.closingdate desc) preclosing 
from t1 a) x 

Select allOpen in new window

In the above code provide by kurt &  rajeevnandanmishra the criteria used is ISNULL. what if instead of ISNULL i want to use '0'

Sorry for the late reply.

Hi,

Below code can be used (or some variation in the HainCurt code, i am not good at CTE):

select x.ledgerid, x.divisionid, x.closingdate, 
	(case when isnull(x.opening,0) <> 0 then x.opening else x.preclosing end ) new_opening, 
	(case when isnull(x.closing,0) <> 0 then x.closing 
		when isnull(x.opening,0) <> 0 then x.opening 
		else preclosing end) new_closing 
from (
select a.*, (select top 1 case when isnull(closing,0) = 0 then opening else closing end from t2 b 
		where a.ledgerid = b.ledgerid and a.divisionid = b.divisionid 
		and b.closingdate < a.closingdate and 
		(isnull(b.closing,0) <> 0 or isnull(b.opening,0)<>0)
		order by b.closingdate desc) preclosing 
from t2 a) x

Select allOpen in new window

But you might need to consider the case if, there is actually opening/closing is going to be zero.

Thanks a lot