Oracle SQL：查找重复的行

Question

I have following table

itemId   Name   PartNum    Price
 1      apple    123       0.99
 2      orange   234       0.5
 3      apple    123       0.99

I want to find the duplicated rows. It should output

ItemId  Name  PartNum  Price
 1     apple    123    0.99
 3     apple    123    0.99

How to do it???????

Answer 1

There are a couple ways to do this basically around joining the table to itself. Here's one solution using a common table expression with the rank() function:

with cte as (
  select itemId, 
    name, 
    partnum, 
    price, 
    rank() over (order by name, partnum, price) rnk
  from yourtable
  ) 
select distinct c.* 
from cte c
  join cte c2 on c.rnk = c2.rnk and c.itemid != c2.itemid

SQL Fiddle Demo

---

Here's an alternative approach:

select distinct y.* 
from yourtable y
  join yourtable y2 on 
      y.name = y2.name and 
      y.partnum = y2.partnum and 
      y.price = y2.price and 
      y.itemid != y2.itemid

More Fiddle

Answer 2

Now I understand and you can do this:

select * from yourTable where name in (
  select name from (
    SELECT Name, PartNum, Price, count(ItemId) qtd
    FROM yourTable 
    group by Name, PartNum, Price,) 
  where qtd>1)

Answer 3

Claudio's answer is pretty close, but to filter the results based on the number of duplicates, you'll want to add a having clause:

select name, partnum, price
from yourTable 
group by name, partnum, price 
having count(itemId) > 1

Answer 4

This is another approach :

Query :

select *
from Table1
where Name||','||PartNum in (
    select Name||','||PartNum
    from Table1
    group by Name, PartNum
    having count(*) > 1)

Results :

| ITEMID |  NAME | PARTNUM | PRICE |
|--------|-------|---------|-------|
|      1 | apple |     123 |  0.99 |
|      3 | apple |     123 |  0.99 |

Answer 5

This is another alternative

SELECT t1.itemId, t1.name, t1.partNum, t1.price
FROM table1 t1
INNER JOIN (SELECT name, partNum, price, COUNT(*) AS count
            FROM table1
            GROUP BY name, partNum, price
            HAVING COUNT(*) > 1
           ) dt ON t1.name = dt.name and t1.partNum = dt.partNum 
                and t1.price = dt.price
ORDER BY t1.itemId

Check it on SQL FIDDLE

Answer 6

Here's another method with fiddle . This uses the analytic function count(*)

select itemid,name,partnum,price
from (
select itemid,
       name,
       partnum,
       price,
       count(*) over (partition by partnum
                      order by price) as part_count
from yourtable
  )
where part_count >1