Question

I would like to have a stored procedure which inserts rows into a table (retrieved from a select query from another table) and for each newly inserted row gets its identity and updates the original table with the identity

Pseudo code-

records = select id,city,state,country from USER where name=@name

for each record in records  // for each rows selected
   insert into LOCATION(city,state,country) values(@record.city,@record.state,@record.country); //inserts a value into LOCATION table
   @id = SCOPE_IDENTITY(); // gets the identity of the newly inserted row
   update USER set LocationId=@id where Id=@record.id //updates the new id back to old table's column
end

This is a data migration task, where we want to segregate the LOCATION from USER table

Thanks in advance for your time and effort for this thread.

Answer 1

You could do something like this:

DECLARE @InsertedValues TABLE (ID INT, City VARCHAR(50), State VARCHAR(50), Country VARCHAR(50))

INSERT INTO dbo.Location(City, State, Country)
  OUTPUT Inserted.ID, Inserted.City, Inserted.State, Inserted.Country INTO @InsertedValues(ID, City, State, Country)
    SELECT City, State, Country
    FROM dbo.YourSourceTable

With this, you now have the inserted values – including the newly defined identity values – in your @InsertedValues table variable and you can now update the source table as you see fit.

UPDATE dbo.YourSourceTable
SET 
   Col1 = iv.Col1,
   Col2 = iv.Col2,  -- etc. - do whatever you nee to do here!
FROM @InsertedValues iv
WHERE  .........  -- here, you need some condition to link the inserted values to the original table

This doesn’t require any cursor or any other messy RBAR (row-by-agonizing-row) processing at all – everything is nicely set-based and as fast as it can possibly be.

Learn more about the OUTPUT clause at MSDN SQL Server Books Online – you can use the OUTPUT clause on insert, update and even delete statements, too!