Question:
How do I count the number of duplicate items in a table?
Answer:
Lets break down your question into several steps.First, let’s create a sample table using the following code:
create table dups( i int)godeclare @i intselect @i = 0while (@i < 35)begin insert into dups(i) values (cast (rand() * 50 as int)) select @i = @i + 1end
Now, let's find rows that are duplicates. For that we can use a simple group by statement:
select i,count(*) as num_recordsfrom dupsgroup by ihaving count(*) > 1
My sample data produced the following result set:
i num_records ----------- ----------- 0 25 218 222 227 231 234 244 249 2
This identifies the rows that have duplicates. But it does not return the total number of duplicates in the table. The first change we must make is to recognize the above rows that show 2 contain only one duplicate.
So we want a query that basically sums up the duplicates from the above query. To do so, we take the previous query and can put that in the from statement as a derived table. We then can use the sum function to create the total for us:
select sum(num_dups)from (select i,count(*)-1 as num_dupsfrom dupsgroup by ihaving count(*)-1 > 0) as mydups