How to join 2 tables without value duplication in PostgreSql - postgresql

Question

I am joining two tables using:
select table1.date, table1.item, table1.qty, table2.anotherQty
from table1
INNER JOIN table2
on table1.date = table2.date
table1
date | item | qty
july1 | itemA | 20
july1 | itemB | 30
july2 | itemA | 20
table2
date | anotherQty
july1 | 200
july2 | 300
Expected result should be:
date | item | qty | anotherQty
july1 | itemA | 20 | 200
july1 | itemB | 30 | null or 0
july2 | itemA | 20 | 300
So that when i sum(anotherQty) it will have 500 only, instead of:
date | item | qty | anotherQty
july1 | itemA | 20 | 200
july1 | itemB | 30 | 200
july2 | itemA | 20 | 300
That is 200+200+300 = 700

Answer 1

SQL DEMO
WITH T1 as (
SELECT *, ROW_NUMBER() OVER (PARTITION BY "date" ORDER BY "item") as rn
FROM Table1
), T2 as (
SELECT *, ROW_NUMBER() OVER (PARTITION BY "date" ORDER BY "anotherQty") as rn
FROM Table2
)
SELECT *
FROM t1
LEFT JOIN t2
ON t1."date" = t2."date"
AND t1.rn = t2.rn
OUTPUT
Filter the columns you want, and change the order if need it.
| date | item | qty | rn | date | anotherQty | rn |
|-------|-------|-----|----|--------|------------|--------|
| july1 | itemA | 20 | 1 | july1 | 200 | 1 |
| july1 | itemB | 30 | 2 | (null) | (null) | (null) |
| july2 | itemA | 20 | 1 | july2 | 300 | 1 |

Answer 2

Try the following code, but know that so long as the qty values differ across rows, that you're going to still get the 'anotherQty' field breaking out into distinct values:
select
table1.date,
table1.item,
table1.qty,
SUM(table2.anotherQty)
from table1
INNER JOIN table2
on table1.date = table2.date
GROUP BY
table1.item,
table1.qty,
table1.date
If you need it to always aggregate down to a single line per item/date, then you will need to add a SUM() to table1.qty as well. Alternately, you could run a common table expression (WITH() statement) for each quantity that you want, summing them within the common table expression, and then rejoining the expressions to your final SELECT statement.
Edit:
Based on the comment from #Juan Carlos Oropeza, I'm not sure that there is a way to get the summed value of 500 while including table1.date in your query, because you will have to group the output by date which will cause the aggregation to split into distinct lines. The following query will get you the sum of anotherQty, at the sacrifice of displaying date:
select
table1.item,
SUM(table1.qty),
SUM(table2.anotherQty)
from table1
INNER JOIN table2
on table1.date = table2.date
GROUP BY
table1.item
If you need date to persist, you can get the sum to show up by using a WINDOW function, but note that this is essentially doing a running sum, and may throw off any subsequent summation you're doing on this query's output in terms of post-processing:
select
table1.item,
table1.date,
SUM(table1.qty),
SUM(table2.anotherQty) OVER (Partition By table1.item)
from table1
INNER JOIN table2
on table1.date = table2.date
GROUP BY
table1.item,
table1.date,
table2.anotherQty