Question

I want to update many rows in a single command with different IDs. These IDs should be passed in the query like

Update table
Set Updated = 1
where ID is in 
(
     1
     2
     3
)

The further difficulty is that there are two primary key columns to check for. I don’t want to update each row in its own command because its very slow because of the server roundtrips.

What is the best way to update a table like this?

Answer 1

Join the table to a derived table that contains a VALUES clause (SQL 2008) or a series of UNION ALL SELECTs.

UPDATE T
SET Col = 1
FROM dbo.Table T
INNER JOIN (
   VALUES
      (1, 3),
      (3, 5),
      (5, 7)
) X (ID1, ID2)
   ON T.ID1 = X.ID1
   AND T.ID2 = X.ID2

For SQL 2005 and earlier:

UPDATE T
SET Col = 1
FROM dbo.Table T
INNER JOIN (
   SELECT 1, 3
   UNION ALL SELECT 3, 5
   UNION ALL SELECT 5, 7
) X (ID1, ID2)
   ON T.ID1 = X.ID1
   AND T.ID2 = X.ID2

Alternately, insert to a temp table and join to that. You could also create a stored procedure that accepts a parameter containing all the keys you need to join with. The parameter could be text (which you would split), xml, or a table-valued parameter.

If the amount of data is simply enormous then you might consider doing a bulk load of the keys from a text file into a table.