Question

I have the following table:

Transaction_ID  int(9)  NO  PRI NULL    auto_increment
Datetime    date    NO      NULL     
Giver_ID    int(9)  NO      NULL     
Recipient_ID    int(9)  NO      NULL     
Points  int(4)  NO      NULL     
Category_ID int(3)  NO      NULL     
Reason  text    NO      NULL     

Apparently I’ve managed to award some of our students double points for “Behaviour Week”.

I would like a list of all the rows in this transactions table where a single user (Recipient_ID) has more than one transaction with Category_ID = 16.

Each Recipient ID should only have ONE transaction with Category_ID = 16.

For example, if I had…

54784   2012-02-01  138573  137444  100 16  You have done very well in Behaviour Week! Because...
55040   2012-02-01  138573  132629  100 16  You have done very well in Behaviour Week! Because...

That’s two separate users, which is fine.

However, if I find…

54784   2012-02-01  138573  137444  100 16  You have done very well in Behaviour Week! Because...
55040   2012-02-01  138573  137444  100 16  You have done very well in Behaviour Week! Because...

Then I’ve made a mistake and I need to delete one of those rows.

Could someone please suggest a query which will list all of these potentially incorrect transactions? I’d prefer for them just to be “SELECT” so that I can manually check rather than blindly DELETE-ing.

Answer 1

To find Recipients IDs which have more than 1 transaction for category 16 you can use the following:

select Recipient_ID,
       Count(Transaction_ID) as [Transactions]
from myTable
where Category_ID = 16
group by Recipient_ID
having Count(Transaction_ID) > 1

If you want to really put your mind at rest, remove the Having clause:

select Recipient_ID,
       Count(Transaction_ID) as [Transactions]
from myTable
where Category_ID = 16
group by Recipient_ID

You can then see the number of transactions against each recipient ID and satisfy yourself that it is really 1 for each