Question

I have a table, tblDocs that has a few columns:

DocName varchar(50)
DocLocation int
Active int
DocID int

All entries in the table have a DocName and DocLocation. Active and DocID are blank.

What I need to do is for each row in tblDocs I need to check the value of DocLocation and based on that value I update tblDocs, setting the Active and DocID columns. I was thinking that I would use a CASE WHEN but I’m not sure – I dont know sql very well. If i were using c# or vb.net it would be the equivalent of a for each loop.

foreach row as DBRow in tblDocs
    row.active = 1
    row.docID = ID
next

How do you do this for SQL Server 2005?

UPDATE

From a couple of responses below it sounds like I will use an UPDATE Statement. So I could accomplish this by doing something like

UPDATE tblDocs
SET docID = 
    CASE DocLocation
        WHEN 1 THEN --do stuff
        WHEN 3 THEN --do other stuff

and this would go through ALL the rows of tblDocs and update the DocID to be based on DocLocation?

Answer 1

If i were using c# or vb.net it would be the equivalent of a for each loop.

The trick here is that C# and VB.Net (and javascript, c++, and any other language you’re likely used to) are procedural languages. SQL is a declarative language. You describe operations that act on an entire set a time.

With that in mind, you want to describe an operation that will "check the value of DocLocation and based on that value I update tblDocs, setting the Active and DocID columns." And to do that you use an UPDATE query and case statements:

UPDATE tblDocs
SET 
  Active = CASE WHEN DocLocation = 'something' THEN 1 ELSE 0 END,
  DocID = CASE WHEN DocLocation = 'something' THEN NULL ELSE DocID END

Just a note that this sample also demonstrates how you can use the CASE statement to unset a value (set it back to null) or even leave it alone by assigning it itself. It’s also worth nothing that if you really expect to leave a lot of the values alone, you should do that with a WHERE clause instead.