Select rows that do not have the opposite value in another row - sql

Question

I am trying to write a query that will only select rows that does not have the opposite value. For example if a column (payments) has 2 negative numbers(-11) and 3 positive numbers(11) both the negative numbers would cancel out and one positive number will remain. I may be explaining this wrong. But any help is appreciated.
table :
CREATE TABLE hamzachecks(
ID VARCHAR(2) NOT NULL
,CHECK VARCHAR(10) NOT NULL
,Payment VARCHAR(8) NOT NULL
);
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('1','9549549544','-112.96');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('2','9549549544','-112.96');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('3','9549549544','112.96');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('4','9549549544','112.96');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('5','9549549544','-165.92');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('6','9549549544','225.92');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('7','9549549544','-299.3');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('8','9549549544','-299.3');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('9','9549549544','-299.3');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('10','9549549544','299.3');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('11','9549549544','299.3');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('12','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('13','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('14','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('15','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('16','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('17','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('18','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('19','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('20','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('21','9549549544','-415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('22','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('23','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('24','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('25','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('26','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('27','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('28','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('29','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('30','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('31','9549549544','415.14');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('32','9549549544','-1024.22');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('33','9549549544','1024.22');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('34','9549549578','-253.77');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('35','9549549578','253.77');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('36','9549549578','-3332.16');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('37','9549549578','-6664.29');
INSERT INTO hamzachecks(ID,CHECK,Payment) VALUES ('38','9549549578','6664.29');

Answer 1

The basic logic is the same as #Marty's, assign a row number to each row in with the same chk/payment combination.
SELECT chk
,id
,Payment
,Row_Number() Over (PARTITION BY chk, Payment ORDER BY Id) AS rn
FROM hamzachecks
order by
chk
-- remove the leading '-'', for numeric data: abs(Payment)
,substring(Payment, charindex('-', Payment)+1, 8000)
,rn
,Payment
It there's a matching row both rows will share the same rn
...
9549549544 3 112.96 1 -- 1st group
9549549544 1 -112.96 1 -- matching value: remove
9549549544 4 112.96 2 -- 2nd group
9549549544 2 -112.96 2 -- matching value: remove
...
9549549544 10 299.3 1 -- 1st group
9549549544 7 -299.3 1 -- matching value: remove
9549549544 11 299.3 2 -- 2nd group
9549549544 8 -299.3 2 -- matching value: remove
9549549544 9 -299.3 3 -- 3rd group, no matching value: keep
...
Now remove the groups with two rows using aggregation:
;WITH cte AS
(
SELECT chk
,id
,Payment
,Row_Number() Over (PARTITION BY chk, Payment ORDER BY Id) AS rn
FROM hamzachecks
)
SELECT
Min(id)
,chk
,max(Payment)
FROM cte
GROUP BY
chk
-- remove the leading '-'', for numeric data: abs(Payment)
,substring(Payment, charindex('-', Payment)+1, 8000)
,rn
HAVING count(*) = 1
See fiddle

Answer 2

I'm going to ignore possible considerations for why you want to do this particular job in this particular way.
So basically: the following SQL will give you the result you are asking for...
;with b (id, payment, rn) as
(select id, payment, row_number() over (partition by Payment order by Id) as rn
from hamzachecks
where payment < 0),
a (id, payment, rn) as
(select id, payment, row_number() over (partition by Payment order by Id) as rn
from hamzachecks
where payment > 0),
matched_ids (id1,id2) as
(select a.id id1, b.id id2
from a
join b on a.payment = -b.payment and a.rn = b.rn)
select *
from hamzachecks
where id not in (select id1 from matched_ids union select id2 from matched_ids)
order by id
This needs a change in your schema: define the Payment column as MONEY (or other number type) instead of VARCHAR(8).
This would be a starting point for further work, as this solution is ignoring the CHECK column.
Explanation: first we split the source table into two groups: a - above zero (payment > 0) and b - below zero (payment < 0). Both of those groups will get another column row_number() ... that get gets each row an "rowID" - based on the position in the particular group (partition by Payment) that basically says: "This is n-th row with the same Payment". The join part then matches all rows that have the same (but opposite) Payment in the other group and also checks the rowID - we now know all rows that have an opposite value. So the main select just ignores those rows and return only those that don't have an opposite value.