MYSQL query - get rows where price is changed

Question

I have two tables viz. order and orderitem. Order table contains date and orderitem table has quantity and price for that order. I want to find the dates where the price has been changed. for example today the price is 10 and after 2 days it become 11 and then remain 11 for next 5 days then again it becomes 10 after 5 days so the result would be 3 dates 10, 11, 10 and the date would be when the price changed

        Date                     price

        '2015-04-29 14:48:23',  15.99
        '2015-04-30 14:20:11',  15.99
        '2015-04-30 21:05:11',  15.99
        '2015-05-04 01:41:12',  15.99
        '2015-07-28 05:52:20',  15.99
        '2015-07-29 16:55:48',  15.99
        '2015-08-01 00:31:53',  15.99
        '2015-08-15 22:41:49',  15.99
        '2015-08-17 02:37:43',  15.99
        '2015-08-17 13:28:03',  15.99
        '2015-08-23 04:16:31',  15.99   
        '2015-08-27 08:43:48',  15.99
        '2015-08-28 18:42:27',  15.99
        '2015-08-29 15:47:17',  15.99
        '2015-08-30 03:33:15',  15.99
        '2015-10-29 03:45:25',  15.99
        '2015-10-29 18:24:20',  15.99
        '2015-10-30 18:17:18',  15.99
        '2015-11-02 15:28:16',  16.99
        '2015-11-03 01:59:03',  16.99
        '2015-11-14 18:22:39',  16.99
        '2015-11-20 02:48:59',  16.99
        '2015-12-16 18:59:54',  16.99
        '2015-12-28 04:08:22',  16.99
        '2016-01-12 03:21:35',  16.99
        '2016-01-18 00:43:56',  16.99
        '2016-01-18 20:11:23',  16.99
        '2016-02-10 19:07:57',  16.99
        '2016-02-26 14:24:29',  16.99
        '2016-03-28 10:17:24',  16.99
        '2016-03-31 23:33:53',  16.99
        '2016-04-01 20:03:03',  17.99
        '2016-05-31 20:30:50',  15.99

I want to get all those date where the price has been changed. Here the output would be

        '2015-04-29 14:48:23',  15.99
        '2015-11-02 15:28:16',  16.99
        '2016-04-01 20:03:03',  17.99
        '2016-05-31 20:30:50',  15.99

i.e. show dates when the price has been changed.

I have written this MYSQL query so far but it gives all rows and does not filter the results:

         SELECT date, price FROM orderitem
         inner join orders o1 on o1.orderid = orderitem.orderid
         where productname = 'IBC'
        and price <> (
        select price from orderitem
        inner join orders on orders.orderid = orderitem.orderid
        where orders.date< o1.date
        order by date desc
        limit 1
        )

Any advise would be appreciated.

Answer 1

For something like this, you need two passes, because you will be comparing each row to the one before it.

One such way of doing this would be to first:

SET @lastprice = 0.0;

Then, your "first pass" will look like:

SELECT `Date`, @lastprice AS `oldprice`, @lastprice := `price` AS `newprice`
FROM `tablename` ORDER BY `Date` ASC;

The order is important here, as you want to build a chronological timeline.

Then, all you need to do is retrieve the rows from this result where the old and new prices are different. Something like this will do:

SELECT * FROM ("Above subquery here") `tmp` WHERE `oldprice` != `newprice`;

Your final query would look something like this:

SET @lastprice = 0.0;
SELECT * FROM (
    SELECT `Date`, @lastprice AS `oldprice`, @lastprice := `price` AS `newprice`
    FORM `tablename` ORDER BY `Date` ASC
) `tmp`
WHERE `oldprice` != `newprice`;

Alternatively, you may prefer to use a temporary table to store the subquery results. In that case, it would look something like this:

SET @lastprice = 0.0;
CREATE TEMPORARY TABLE `tmp`
    SELECT `Date`, @lastprice AS `oldprice`, @lastprice := `price` AS `newprice`
    FROM `tablename` ORDER BY `Date` ASC;
SELECT * FROM `tmp` WHERE `oldprice` != `newprice`;
DROP TEMPORARY TABLE `tmp`;

The exact implementation is up to you, but this would be a good way to do it.