Nearest value

Question

i have a target of developing aging report of vendor payments
but payment is not marking billwise
now if party credit amount is 50000
then i have to go to the nearest record to 50000
for e.g there are eleven bills of party
last one amount is 15000
seocnd last 16000
third last 17000
and then 40000
the need a query that will fetch records upto 17000 as the sum of last three payments equals to (48000) that is nearest to 50000 of total credit

Answer 1

I'm not sure what you're asking there?
Do you mean what happens if the latest and latest but one are equally close to the specified amount (one above, one below)?
I've gone with including the amount above the cutoff limit (figuring that if these are bills that the company demands the debtor must pay, then the company wants to get as much money as possible).
changing
where   abs_diff <= nvl(abs_diff_prev, abs_diff)
to
where   abs_diff < nvl(abs_diff_prev, abs_diff+1)works if you only want to include the row less than the set limit in that specific case.
Eg:
with my_tab as (select 1 cust_id, 40000 amount, trunc(sysdate -19) bill_date from dual union all
                select 1 cust_id, 17000 amount, trunc(sysdate -16) bill_date from dual union all
                select 1 cust_id, 16000 amount, trunc(sysdate -5) bill_date from dual union all
                select 1 cust_id, 15000 amount, trunc(sysdate -1) bill_date from dual union all
                select 1 cust_id, 22000 amount, trunc(sysdate -22) bill_date from dual union all
                select 2 cust_id, 20000 amount, trunc(sysdate -2) bill_date from dual union all
                select 2 cust_id, 29000 amount, trunc(sysdate -1) bill_date from dual union all
                select 2 cust_id, 2000 amount, trunc(sysdate -6) bill_date from dual union all
                select 2 cust_id, 14000 amount, trunc(sysdate -10) bill_date from dual union all
                select 3 cust_id, 20000 amount, trunc(sysdate -2) bill_date from dual union all
                select 3 cust_id, 31000 amount, trunc(sysdate -8) bill_date from dual union all
                select 3 cust_id, 12000 amount, trunc(sysdate -10) bill_date from dual union all
                select 4 cust_id, 66000 amount, trunc(sysdate -2) bill_date from dual)
-- end of creating mock-up of my_tab table. Now, onto the query:
select cust_id, amount, bill_date
from   (select cust_id, amount, bill_date, abs_diff, lag(abs_diff) over (partition by cust_id
                                                                          order by bill_date desc) abs_diff_prev
        from   (select cust_id,
                       amount,
                       bill_date,
                       abs(50000-sum(amount) over (partition by cust_id
                                         order by bill_date desc)) abs_diff
                from   my_tab))
where   abs_diff <= nvl(abs_diff_prev, abs_diff+1)
order by cust_id, bill_date desc
CUST_ID     AMOUNT     BILL_DATE
1     15000     05/11/2008
1     16000     01/11/2008
1     17000     21/10/2008
2     29000     05/11/2008
2     20000     04/11/2008
2     2000     31/10/2008
3     20000     04/11/2008
3     31000     29/10/2008
4     66000     04/11/2008
with my_tab as (select 1 cust_id, 40000 amount, trunc(sysdate -19) bill_date from dual union all
                select 1 cust_id, 17000 amount, trunc(sysdate -16) bill_date from dual union all
                select 1 cust_id, 16000 amount, trunc(sysdate -5) bill_date from dual union all
                select 1 cust_id, 15000 amount, trunc(sysdate -1) bill_date from dual union all
                select 1 cust_id, 22000 amount, trunc(sysdate -22) bill_date from dual union all
                select 2 cust_id, 20000 amount, trunc(sysdate -2) bill_date from dual union all
                select 2 cust_id, 29000 amount, trunc(sysdate -1) bill_date from dual union all
                select 2 cust_id, 2000 amount, trunc(sysdate -6) bill_date from dual union all
                select 2 cust_id, 14000 amount, trunc(sysdate -10) bill_date from dual union all
                select 3 cust_id, 20000 amount, trunc(sysdate -2) bill_date from dual union all
                select 3 cust_id, 31000 amount, trunc(sysdate -8) bill_date from dual union all
                select 3 cust_id, 12000 amount, trunc(sysdate -10) bill_date from dual union all
                select 4 cust_id, 66000 amount, trunc(sysdate -2) bill_date from dual)
-- end of creating mock-up of my_tab table. Now, onto the query:
select cust_id, amount, bill_date
from   (select cust_id, amount, bill_date, abs_diff, lag(abs_diff) over (partition by cust_id
                                                                          order by bill_date desc) abs_diff_prev
        from   (select cust_id,
                       amount,
                       bill_date,
                       abs(50000-sum(amount) over (partition by cust_id
                                         order by bill_date desc)) abs_diff
                from   my_tab))
where   abs_diff < nvl(abs_diff_prev, abs_diff+1)
order by cust_id, bill_date desc
CUST_ID     AMOUNT     BILL_DATE
1     15000     05/11/2008
1     16000     01/11/2008
1     17000     21/10/2008
2     29000     05/11/2008
2     20000     04/11/2008
3     20000     04/11/2008
3     31000     29/10/2008
4     66000     04/11/2008Note the difference between the two results for cust_id = 2.