Find duplicates in SQL table, list multiple instances - sql

Question

I'm trying to list the multiple instances of tbldoc.[docid] from tbldoc where tbldoc.[filename] occurs more than once, id like them seperated by comma and grouped by [filename]
this code works great to find duplicates:
SELECT cast([filename] as varchar(max)),
COUNT(cast([filename] as varchar(max)))
FROM tbldoc
GROUP BY cast([filename] as varchar(max))
HAVING ( COUNT(cast([filename] as varchar(max))) > 1 )
but when i try adding [docid] i get an error:
Column 'tbldoc.DocID' is invalid in the select list because it is not
contained in either an aggregate function or the GROUP BY clause.
this is what i am trying:
SELECT [docid], cast([filename] as varchar(max)),
COUNT(cast([filename] as varchar(max)))
FROM tbldoc
GROUP BY cast([filename] as varchar(max))
HAVING ( COUNT(cast([filename] as varchar(max))) > 1 )
I have no idea how to get all of the [docid]s to list seperated by commas, I'm a a pretty new user when it comes to sql.
this is the output i would like to see:
[docids]|[filemame]|[instances]
12345,12346| excelfile.xls | 3
Thanks ahead of time for the help guys/gals! =)

Answer 1

Iyosha,
You need to join your first result set back to your full table to get the DocIDs. I'll take the CAST() as read to save some typing.
;with CountedFiles as
(
SELECT
filename,
COUNT(filename) as Total
FROM tbldoc
GROUP BY filename
HAVING COUNT(filename) > 1
)
select
cf.filename,
cf.Total,
td.DocID
from CountedFiles as cf
inner join tbldoc at td
on td.filename = cf.filename;
This will return one DocId, one filename and the count per row. You can then follow Adam's link to turn this into a comma list.