Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1286
  • Last Modified:

Problem deleting duplicates using a CTE when there are table joins

I've used CTEs to remove duplicates successfully before, but this time round I wanted to delete them from one table having joined to other tables and it comes back with the error:

View or function 'dupnotesCTE' is not updatable because the modification affects multiple base tables. See the attached code for my sql statement.

Not sure where I've gone wrong here...can someone help please?
;WITH dupnotesCTE([FileNo]
      ,[CommentDate]
      ,[CommentTime]
      ,[CommentUser]
      ,[Cat]
      ,[Pri]
      ,[Line]
      ,[Note], Ranking)
AS
(
SELECT	p.[FileNo]
      ,p.[CommentDate]
      ,p.[CommentTime]
      ,p.[CommentUser]
      ,p.[Cat]
      ,p.[Pri]
      ,p.[Line]
      ,p.[Note]
	,Ranking = row_number() OVER(PARTITION BY p.[FileNo]
      ,p.[CommentDate]
      ,p.[CommentTime]
      ,p.[CommentUser]
      ,p.[Cat]
      ,p.[Pri]
      ,p.[Line]
      ,p.[Note] ORDER BY p.[FileNo]
      ,p.[CommentDate]
      ,p.[CommentTime]
      ,p.[CommentUser]
      ,p.[Cat]
      ,p.[Pri]
      ,p.[Line]
      ,p.[Note])
FROM filenotes p
join detailtable d on ltrim(rtrim(p.FileNo)) = ltrim(rtrim(d.FileNo))
join country c on d.country_no = c.country_no
where c.continent in ('Asia','Europe')
)

delete from dupnotesCTE
where ranking > 1

Open in new window

0
Grffster
Asked:
Grffster
  • 5
  • 5
1 Solution
 
Erick37Commented:


...

delete p
FROM filenotes p
JOIN dupnotesCTE c ON c.FileNo = p.fileNo --??
where c.ranking > 1
0
 
GrffsterAuthor Commented:
This won't work though will it as, say you have a pair of duplicates A and B. You have A in dupnotesCTE with rank 1, and B in dupnotesCTE with rank 2. When you link from filenotes to dupnotes you'll have 4 records:

filenotes.A and dupnotes.A (rank 1)
filenotes.A and dupnotes.B (rank 2)
filenotes.B and dupnotes.A (rank 1)
filenotes.B and dupnotes.B (rank 2)

Your "where c.ranking > 1" will remove the 2nd and 4th combinations but you're left with two records i.e. combinations 1 and 3 above.

May be mistaken, but my query seems to reflect this.
0
 
Erick37Commented:
Does filenotes table have an identity column or primary key that you could use to join the cte?

0
Technology Partners: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

 
GrffsterAuthor Commented:
No it doesn't sorry.
0
 
Erick37Commented:
So all of the columns with the duplicate FileNo are also duplicate?
0
 
GrffsterAuthor Commented:
Yes, all the fields per fileno duplicate are identical i.e. a true duplicated record with the same data in each column:

[FileNo],[CommentDate],[CommentTime],[CommentUser],[Cat],[Pri],[Line],[Note]
0
 
Erick37Commented:
Having an identity column would greatly simplify identifying the duplicates.

One possibility would be to copy the records over to a temp table with an identity column.
Remove the duplicates from the temp table.
Delete all the records from the user table that you originally copied to the temp table.
Insert the cleaned records back into filenotes table.

Here are some other alternatives:
"How to delete duplicate records or rows among identical rows in a table where no primary key exists"
http://www.kodyaz.com/articles/delete-duplicate-records-rows-in-a-table.aspx
0
 
GrffsterAuthor Commented:
I've added an identity column to the filenotes table now. How should it look now? Do I do something like at the bottom of your last link in method 3?
0
 
Erick37Commented:
You will add the identity column to the CTE and use that to join the table to the cte.

;WITH dupnotesCTE AS
(
  SELECT
      IdentityColumn
      ,[FileNo]
      ...
      ,row_number() OVER(PARTITION BY p.[FileNo] order by IdentityColumn) AS Ranking
     ...

)
delete p
FROM filenotes p
JOIN dupnotesCTE c ON c.IdentityColumn = p.IdentityColumn AND c.ranking > 1
0
 
GrffsterAuthor Commented:
That's worked a dream thanks Eric.
0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

  • 5
  • 5
Tackle projects and never again get stuck behind a technical roadblock.
Join Now