Finding duplicate differences between two tables in sql - sql

Question

I try to find duplicate rows between two tables. This code works only if records are not duplicated:
(select [Name], [Age] from PeopleA
except
select [Name], [Age] from PeopleB)
union all
(select [Name], [Age] from PeopleB
except
select [Name], [Age] from PeopleA)
How to find missing, duplicate records. Robert 34 in PersonA table for example below:
PersonA:
Name | Age
-------------
John | 45
Robert | 34
Adam | 26
Robert | 34
PersonB:
Name | Age
-------------
John | 45
Robert | 34
Adam | 26

Answer 1

You can use UNION ALL to concat both tables and Group By with Having clause to find duplicates:
SELECT x.Name, x.Age, Cnt = Count(*)
FROM (
SELECT a.Name, a.Age
FROM PersonA a
UNION ALL
SELECT b.Name, b.Age
FROM PersonB b
) x
GROUP BY x.Name, x.Age
HAVING COUNT(*) > 1
According to your clarification in the comment, you could use following query to find all name-age combinations in PersonA which are different in PersonB:
WITH A AS(
SELECT a.Name, a.Age, cnt = count(*)
FROM PersonA a
GROUP BY a.Name, a.Age
),
B AS(
SELECT b.Name, b.Age, cnt = count(*)
FROM PersonB b
GROUP BY b.Name, b.Age
)
SELECT a.Name, a.Age
FROM A a LEFT OUTER JOIN B b
ON a.Name = b.Name AND a.Age = b.Age
WHERE a.cnt <> ISNULL(b.cnt, 0)
Demo
If you also want to find persons which are in PersonB but not in PersonA you should use a FULL OUTER JOIN as Gordon Linoff has commented:
WITH A AS(
SELECT a.Name, a.Age, cnt = count(*)
FROM PersonA a
GROUP BY a.Name, a.Age
),
B AS(
SELECT b.Name, b.Age, cnt = count(*)
FROM PersonB b
GROUP BY b.Name, b.Age
)
SELECT Name = ISNULL(a.Name, b.Name), Age = ISNULL(a.Age, b.Age)
FROM A a FULL OUTER JOIN B b
ON a.Name = b.Name AND a.Age = b.Age
WHERE ISNULL(a.cnt, 0) <> ISNULL(b.cnt, 0)
Demo

Answer 2

I like Tim's answer but you need to check in both tables if the records are missing. He is only checking if the records are missing in table A. Try this to check if records are missing in either of the tables and how many times.
Select *, 'PersonB' MissingInTable, a.cnt - isnull(b.cnt,0) TimesMissing From
(
Select *, count(1) cnt from PersonA group by Name, Age) A Left join
(Select *, count(1) cnt from PersonB group by Name, Age) B
On a.age=b.age and a.name=b.name
where a.cnt>isnull(b.cnt,0)
Union All
Select *, 'PersonA' MissingInTable, b.cnt - isnull(a.cnt,0) TimesMissing From
(
Select *, count(1) cnt from PersonA group by Name, Age) A Right join
(Select *, count(1) cnt from PersonB group by Name, Age) B
On a.age=b.age and a.name=b.name
where b.cnt>isnull(a.cnt,0)
See demo here : http://sqlfiddle.com/#!6/06020/13

Answer 3

Add another UNION ALL!
Code:
(SELECT [Name], [Age], 'Missing from B' AS [Type] from PeopleA
EXCEPT
SELECT [Name], [Age], 'Missing from B' AS [Type] from PeopleB)
UNION ALL
(SELECT [Name], [Age], 'Missing from A' as [Type] from PeopleB
EXCEPT
SELECT [Name], [Age], 'Missing from A' AS [Type] from PeopleA)
UNION ALL
SELECT [Name], [Age], 'Duplicate' AS [Type] FROM PeopleA INNER JOIN PeopleB ON PeopleA.Name = PeopleB.Name AND
PeopleA.Age=PeopleB.Age