How can I solve this query in sql server? - sql

Question

Sql Fiddle example
I have this table structure:
CREATE TABLE IF NOT EXISTS `client` (
`id` int(6) unsigned NOT NULL,
`name` varchar(200),
`balance` decimal NOT NULL,
PRIMARY KEY (`id`)
) DEFAULT CHARSET=utf8;
INSERT INTO `client` (`id`, `name`, `balance`) VALUES
('1', 'Pepito', '500');
CREATE TABLE IF NOT EXISTS `balance_Movements` (
`id` int(6) unsigned NOT NULL,
`clientId` int(6),
`movementType` varchar(200),
`import` decimal NOT NULL,
`dateMovement` datetime,
PRIMARY KEY (`id`)
) DEFAULT CHARSET=utf8;
INSERT INTO `balance_Movements` (`id`, `clientid`, `movementType`, `import`, `dateMovement`) VALUES
('1', '1', 'payment', '50', '2018/05/11'),
('2', '1', 'refund', '25.05', '2018/05/10'),
('3', '1', 'refund', '60', '2018/04/06'),
('4', '1', 'payment', '100', '2018/04/03');
Client has at the start : 500€ --> so the declare variable will be like this:
declare #total_balance as decimal;
set #total_balance = (select balance from client where id = 1);
Result
------
500
I need to change after the result of #total_balance, taking the value for the last row:
Example:
Table_balance_Movements
------------------------
Total
-----
450
475.05
535.05
435.05
Explanation:
450 --> The client start with 500, so if the type of movement is payment I need to substract the balance of client to the import movement, and save the data in #total_balance to use it later = 500 - 50 = 450
475.05 --> I get the value for #total_balance and sum 25.05 because in this line the type of movement is refund = 450 + 25.05 = 475.05
535.05 --> the same thing, I get the value from the #total_balance variable and see what type of movement is and substract or sum the import = 475.05 + 60 = 535.05
435.05 --> 535.05 - 100 = 435.05
I want to do something similar to this concept:
declare #total_balance as decimal;
set #total_balance = (select balance from client where id = 1);
select (case when movementType = 'payment' then (#total_balance = #total_balance - import)
when movementType = 'refund' then (#total_balance = #total_balance + import) end) as total
from balance_Movements;
it would be possible? thanks

Answer 1

You can solve this with the cumulative sum window functions. The following will work in SQL Server and MySQL 8+
select c.*,
(c.balance +
sum(case when bm.movement_type = 'payment' then - import
when bm.movement_type = 'refund' then import
end) over (partition by c.id order by bm.datemovement)
) as net_balance
from client c join
balance_movements bm
on bm.clientid = c.id

Answer 2

I think the simplest way is using SUM with condition aggregate function window function to get accumulate the add with balance
SELECT c.balance +SUM(CASE WHEN movementType = 'payment' THEN - import
WHEN movementType = 'refund' THEN import
ELSE 0 END
) OVER(ORDER BY b.id) Total
FROM client c
JOIN balance_Movements b on c.id = b.clientid
But if your dbms didn't support window function, you can try to use correctly subquery in select
SELECT c.balance +
(
SELECT SUM(CASE WHEN movementType = 'payment' THEN - import
WHEN movementType = 'refund' THEN import
ELSE 0 END)
FROM balance_Movements bb
WHERE bb.id <= b.id and b.clientid = bb.clientid
) Total
FROM client c
JOIN balance_Movements b on c.id = b.clientid
sqlfiddle

Answer 3

You could just add the sum of refunds and subtract the sum of payments like this:
SELECT 500 +
(SELECT SUM(import) FROM balance_Movements WHERE movementType = 'refund')
- (SELECT SUM(import) FROM balance_Movements WHERE movementType = 'payment') AS total
Result: 435.05

Answer 4

Give this a shot...
IF OBJECT_ID('tempdb..#balance_Movements', 'U') IS NOT NULL
BEGIN DROP TABLE #balance_Movements; END;
CREATE TABLE #balance_Movements (
id INT NOT NULL,
clientId INT,
movementType VARCHAR (200),
import DECIMAL (9, 2) NOT NULL,
dateMovement DATETIME,
PRIMARY KEY (id)
);
INSERT INTO #balance_Movements (id, clientid, movementType, import, dateMovement) VALUES
('1', '1', 'payment', '50', '2018/05/11'),
('2', '1', 'refund', '25.05', '2018/05/10'),
('3', '1', 'refund', '60', '2018/04/06'),
('4', '1', 'payment', '100', '2018/04/03');
--=============================================================
DECLARE #_start DECIMAL(9,2) = 500;
SELECT
*,
running_total = #_start - SUM(CASE WHEN bm.movementType = 'refund' THEN -1 * bm.import ELSE bm.import END) OVER (PARTITION BY bm.clientId ORDER BY bm.dateMovement desc)
FROM
#balance_Movements bm;
Results...
id clientId movementType import dateMovement running_total
----------- ----------- ------------ ---------- ----------------------- --------------
1 1 payment 50.00 2018-05-11 00:00:00.000 450.00
2 1 refund 25.05 2018-05-10 00:00:00.000 475.05
3 1 refund 60.00 2018-04-06 00:00:00.000 535.05
4 1 payment 100.00 2018-04-03 00:00:00.000 435.05