Question

This is my situation.

TABLE1:

DOCUMENT_ID,
GUID

TABLE2:

DOCUMENT_ID,
FILE

The tables are joined by DOCUMENT_ID, meaning that TABLE2 can have one or many rows with the same DOCUMENT_ID.

My problem is that TABLE2 values for whole bunch of DOCUMENT_ID have same FILE values.

I need a SQL query that will get me all GUID and count how many rows in TABLE2 for this DOCUMENT_ID have EXACTLY THE SAME FILE value (so that I can copy the GUID to Excel).

Then I need to UPDATE TABLE2‘s FILE columns for these cases.

For instance if DOCUMENT_ID has three rows in TABLE2 with same FILE value, I need to update two of them by adding a postfix like FILEVALUE-1, FILEVALUE-2 and so on.

Hope I make sense.

To all experts thank you in advance.

Answer 1

To get duplicates you might employ oldfashioned group by:

select table1.guid, table1.document_id, table2.[file], count(*) cnt
  from table1
 inner join table2
    on table1.document_id = table2.document_id
 group by table1.guid, table1.document_id, table2.[file]
having count (*) > 1

To directly update duplicates, you might use CTE:

; with t2 as (
    select id, 
           [file],
           row_number() over (partition by document_id, [file]
                              order by id) rn
      from table2
)
update t2
   set [file] = [file] + '-' + convert(varchar(10), rn - 1)
 where t2.rn > 1

Note that I’ve added ID as a placeholder for primary key. You need a way to identify a record to be updated.

There is live test @ Sql Fiddle.