CODEWITHSUNDEEP
sqloracle-databasedatecount
Jama
Jama

Reputation: 295

SQL - Count number of changes in an ordered list

Say I've got a table with two columns (date and price). If I select over a range of dates, then is there a way to count the number of price changes over time?

For instance:

   Date   | Price
22-Oct-11 |  3.20
23-Oct-11 |  3.40
24-Oct-11 |  3.40
25-Oct-11 |  3.50
26-Oct-11 |  3.40
27-Oct-11 |  3.20
28-Oct-11 |  3.20

In this case, I would like it to return a count of 4 price changes.

Thanks in advance.

Upvotes: 3

Views: 2862

Answers (3)

Yahia
Yahia

Reputation: 70379

Depending on the Oracle version use either analytical functions (see answer from Justin Cave) or this

SELECT 
SUM (CASE WHEN PREVPRICE != PRICE THEN 1 ELSE 0 END) CNTCHANGES
FROM
(
SELECT
C.DATE,
C.PRICE,
MAX ( D.PRICE ) PREVPRICE
FROM
(
SELECT 
A.Date, 
A.Price,
(SELECT MAX (B.DATE) FROM MyTable B WHERE B.DATE < A.DATE) PrevDate
FROM MyTable A
WHERE A.DATE BETWEEN YourStartDate AND YourEndDate
) C
INNER JOIN MyTable D ON D.DATE = C.PREVDATE
GROUP BY C.DATE, C.PRICE
)

Upvotes: 0

Justin Cave
Justin Cave

Reputation: 231851

You can use the analytic functions LEAD and LAG to access to prior and next row of a result set and then use that to see if there are changes.

SQL> ed
Wrote file afiedt.buf

  1  with t as (
  2    select date '2011-10-22' dt, 3.2 price from dual union all
  3    select date '2011-10-23', 3.4 from dual union all
  4    select date '2011-10-24', 3.4 from dual union all
  5    select date '2011-10-25', 3.5 from dual union all
  6    select date '2011-10-26', 3.4 from dual union all
  7    select date '2011-10-27', 3.2 from dual union all
  8    select date '2011-10-28', 3.2 from dual
  9  )
 10  select sum(is_change)
 11    from (
 12      select dt,
 13             price,
 14             lag(price) over (order by dt) prior_price,
 15             (case when lag(price) over (order by dt) != price
 16                   then 1
 17                   else 0
 18               end) is_change
 19*       from t)
SQL> /

SUM(IS_CHANGE)
--------------
             4

Upvotes: 5

Sparky
Sparky

Reputation: 15115

Try this

select count(*)
from 
(select date,price from table where date between X and Y
 group by date,price )

Upvotes: 0

Related Questions