Question

I have a table which stores monthly billing information.

CREATE TABLE [dbo].[billing_history](
[id] [numeric](18, 0) IDENTITY(1,1) NOT NULL,
[reading_date] [date] NOT NULL,
[reading] [numeric](18, 0) NOT NULL,
[consumer_id] [int] NOT NULL)

The consumer_id is a foreign key referencing the consumer details table.

What i want is to subtract every customer current reading from the reading of the previous month. This would generate the current bill. Any ideas.

Answer 1

You could use something similar to this, where you would replace the values of the month/year that you want to return:

select b1.consumer_id,
  sum(b1.reading - isnull(b2.reading, 0)) Total
from billing_history b1
left join billing_history b2
  on b1.consumer_id = b2.consumer_id
  and month(b2.reading_date) =12
  and year(b2.reading_date) = 2012
where month(b1.reading_date) = 1
  and year(b1.reading_date) = 2013
group by b1.consumer_id;

See SQL Fiddle with Demo.

If you don’t want to pass in the values of the month and year to search and you only want the current/previous month, then you could use something similar to this using a CTE:

;with cur as
(
  select consumer_id,
    reading,
    month(getdate()) curMonth,
    year(getdate()) curYear,
    case when month(getdate()) = 1 then 12 else month(getdate()) -1 end preMonth,
    case when month(getdate()) = 1 then year(getdate())-1 else year(getdate()) end preYear
  from billing_history
  where month(reading_date) = month(getdate())
    and year(reading_date) = year(getdate())
)
select c.consumer_id, 
  sum(c.reading - isnull(pre.reading, 0)) TotalReading
from cur c
left join billing_history pre
  on c.consumer_id = pre.consumer_id
  and month(pre.reading_date) = c.preMonth
  and year(pre.reading_date) = c.preYear
group by c.consumer_id

See SQL Fiddle with Demo

This version gets both the current/previous month and year values to be used. If you are not familiar with CTE syntax, this can also be written as:

select c.consumer_id, 
  sum(c.reading - isnull(pre.reading, 0)) TotalReading
from
(
  select consumer_id,
    reading,
    month(getdate()) curMonth,
    year(getdate()) curYear,
    case when month(getdate()) = 1 then 12 else month(getdate()) -1 end preMonth,
    case when month(getdate()) = 1 then year(getdate())-1 else year(getdate()) end preYear
  from billing_history
  where month(reading_date) = month(getdate())
    and year(reading_date) = year(getdate())
) c
left join billing_history pre
  on c.consumer_id = pre.consumer_id
  and month(pre.reading_date) = c.preMonth
  and year(pre.reading_date) = c.preYear
group by c.consumer_id;

See SQL Fiddle with Demo.

As you can see in my queries I used the aggregate function SUM() and a GROUP BY on the consumer_id. I did this in the event you had more than one reading per customer. If you know you will only have one reading per month, then you could remove the aggregate.