MS Access SQL join only display extra records - sql

Question

I'm trying to display left over records after matching one-to-one rows. How do I display extra/left over records after joining two tables?
Suppose I have two tables, A and B. They both display the the same transactions at the end of the day. However, Table A has more detail about the records but is late in getting updated. Table B, on the other hand, has limited information about transactions but is updated several hours before Table A.
I need a query that can return which records have yet to appear in Table A from Table B.
TABLE A
+-------+-----+---------+----------+---------------------------+
| NAME | ID | AMOUNT | TYPE | PROCESSED TIMESTAMP |
+-------+-----+---------+----------+---------------------------+
| ABC | 123 | -420.07 | PURCHASE | 2018-09-06-08.26.32.000000|
| ABC | 123 | 420.07 | REFUND | 2018-09-06-07.12.18.000000|
| BBC | 456 | -5.00 | PURCHASE | 2018-09-06-10.25.13.000000|
+-------+-----+---------+----------+---------------------------+
TABLE B
+----+----------+---------------------------+
| ID | AMOUNT | RECEIVED TIMESTAMP |
+----+----------+---------------------------+
|123 | -420.07 | 2018-09-05-09.26.15.000000|
|123 | 420.07 | 2018-09-05-08.12.03.000000|
|123 | -420.07 | 2018-09-05-08.40.00.000000|
|456 | -5.00 | 2018-09-05-08.45.00.000000|
+----+----------+---------------------------+
QUERY RESULTS
+----+----------+
| ID | AMOUNT |
+----+----------+
|123 | -420.07 |
+----+----------+
I can manage to find all the records related to the ID that is "off balance" but I need only the specific records that are extra:
SELECT * FROM b
WHERE id
IN
(SELECT d.id AS id
FROM
(SELECT * FROM
(SELECT id, ROUND(SUM(amount),2) AS balance FROM a GROUP BY id) c
RIGHT JOIN
(SELECT id, ROUND(SUM(amount),2) AS balance FROM b GROUP BY id) d
ON c.id = d.id
WHERE c.balance <> d.balance))
Yields...
+----+----------+
| ID | AMOUNT |
+----+----------+
|123 | -420.07 |
|123 | 420.07 |
|123 | -420.07 |
+----+----------+

Answer 1

You need to read up more on joins . There are 3 basic joins which can make life much simpler.
INNER JOIN: First, this is not asked, but the query you have provided for finding off balance items is too complicated. It can be simplified by an inner join.
Inner join is a set operation which will basically get data from both tables (set) which match the condition.
select * from
(
(select id, sum(amount) from a group by id) group_A
INNER JOIN (select id, sum(amount) from b group by id) group_B
ON group_A.id = group_B.id
WHERE group_A.balance != group_B.balance
)
LEFT/RIGHT OUTER JOIN: Left outer join is an operation which will return all the data that is present in both sets and also the data that is in left set but not the right set. Right join is essential same operation on the right set. Important to notice that the extra records pulled here would be null for the side where they do not exist.
Since you want records which are present in table B but not in A, there are multiple ways of achieving this, one would be to get records present in both tables (inner join) and then get all the records in table B but not in the inner join done earlier. Using definition of group_A/group_B from the inner join example.
select id from b where id not in (
select id from group_A INNER JOIN group_B on group_A.id = group_B.id)
Or you can do a right outer join and then using the property of that fields fetched from table A would be coming as null, can filter out the required ID.
select group_B.id from group_A RIGHT OUTER JOIN group_B ON group_A.id = group_B.id
where group_A.id is null
Please use the primary keys on the joins to get the correct results as mentioned by user #ComputerVersteher

Answer 2

I think, you should add PK col.
I can't match data with table A and B, and can't seperate 2 rows at table B.
+----+----------+---------------------------+
| ID | AMOUNT | RECEIVED TIMESTAMP |
+----+----------+---------------------------+
|123 | -420.07 | 2018-09-05-09.26.15.000000|<-
|123 | 420.07 | 2018-09-05-08.12.03.000000|
|123 | -420.07 | 2018-09-05-08.40.00.000000|<-
|456 | -5.00 | 2018-09-05-08.45.00.000000|
+----+----------+---------------------------+
I add new col(deal_no) and made it.
https://www.db-fiddle.com/f/3GfZoQwGhBLf7YWf2RucBF/4
select tmp_B.deal_no, tmp_B.id, tmp_B.amount, tmp_A.deal_no
from tmp_B
left outer join tmp_A
on tmp_A.deal_no = tmp_B.deal_no
where tmp_A.deal_no is null