SSIS Update Table Values from Another Table - sql-server

Question

I want to update TableA with values from TableB on a nightly basis. Right now, I'm trying to do this with SSIS 2005 using a Script Task with the SQL in it. Each time I try to do the update in this manner, I get a time-out error.
Is there a better way to do this in SSIS?
Current information below:
Public Sub Main()
Const Component_Name As String = "Updating TableA Data"
Const Conn_String As String = "Data Source=DB_A;Initial Catalog=TableA;Integrated Security=SSPI;"
Const sql_Emp As String = "Update TableA Set Contract = c.License_No, SEIN = convert(varchar, c.Lic_Exp_Date, 101) " _
& "From Server.DB_B.dbo.TableB c Inner Join TableA b on " _
& "rtrim(ltrim(c.business_lic)) = rtrim(ltrim(cast(b.Account_Key as varchar(14)))) " _
& "Where c.Lic_Exp_Date = (select Max(Lic_Exp_Date) From Server.DB_B.dbo.TableB " _
& "Where rtrim(ltrim(business_lic)) = rtrim(ltrim(cast(b.Account_Key as varchar(14))))) " _
& "and convert(varchar, c.Lic_Exp_Date, 101) <> convert(varchar, b.SEIN, 101)"
Dim con As SqlConnection = New SqlConnection(Conn_String)
Try
Dts.Log("Opening DB Connection: " & con.ConnectionString, 0, Nothing)
con.Open()
Dim duh As New SqlCommand(sql_Emp, con)
duh.ExecuteNonQuery()
con.Close()
Dts.Log(String.Format(Component_Name), 0, Nothing)
Dts.Events.FireInformation(0, Component_Name, String.Format("TableA Data Updating"), "", 0, True)
Dts.TaskResult = Dts.Results.Success
Catch ex As Exception
Dts.Events.FireError(0, Component_Name, ex.Message, "", 0)
Dts.Log("Exception detected: " & ex.ToString, 0, Nothing)
Dts.TaskResult = Results.Failure
End Try
End Sub

Answer 1

Let's start by cleaning that up a little:
Public Sub Main()
Const Component_Name As String = "Updating TableA Data"
Const Conn_String As String = "Data Source=DB_A;Initial Catalog=TableA;Integrated Security=SSPI;"
Const sql_Emp As String = _
"UPDATE TableA" _
+ " SET Contract = c.License_No, SEIN = convert(varchar, c.Lic_Exp_Date, 101)" _
+ " FROM Server.DB_B.dbo.TableB c" _
+ " INNER JOIN TableA b" _
+ " ON rtrim(ltrim(c.business_lic)) = rtrim(ltrim(cast(b.Account_Key as varchar(14))))" _
+ " WHERE c.Lic_Exp_Date= (" _
+ " SELECT MAX(Lic_Exp_Date)" _
+ " FROM Server.DB_B.dbo.TableB" _
+ " WHERE rtrim(ltrim(business_lic)) = rtrim(ltrim(cast(b.Account_Key as varchar(14))))" _
+ ") AND convert(varchar, c.Lic_Exp_Date, 101) <> convert(varchar, b.SEIN, 101)"
Try
Using con As New SqlConnection(Conn_String), _
cmds New SqlCommand(sql_Emp, con)
Dts.Log("Opening DB Connection: " & con.ConnectionString, 0, Nothing)
con.Open()
cmd.ExecuteNonQuery()
Dts.Log(String.Format(Component_Name), 0, Nothing)
Dts.Events.FireInformation(0, Component_Name, String.Format("TableA Data Updating"), "", 0, True)
Dts.TaskResult = Dts.Results.Success
End Using
Catch ex As Exception
Dts.Events.FireError(0, Component_Name, ex.Message, "", 0)
Dts.Log("Exception detected: " & ex.ToString, 0, Nothing)
Dts.TaskResult = Results.Failure
End Try
End Sub
Okay, now that I can read it I can start looking at what might be broken. Check back in a few minutes for edits.
Okay, now let's look at that query. I'm missing some data type information, so I'm going to make some assumptions. Please correct any that are wrong:
b.Account_Key is some number type, probably int. Otherwise you wouldn't need to convert to varchar
The Lic_Exp_Date columns really are of type datetime
If those are correct, I think this will do what you want, but do it a lot faster:
UPDATE TableA
SET Contract = c1.License_No, SEIN = DATEADD(dd,0, DATEDIFF(dd,0, c1.Lic_Exp_Date))
FROM TableA b
INNER JOIN Server.DB_B.dbo.TableB c1
ON ISNUMERIC(c1.busines_lic) = 1 AND cast(c1.business_lic AS int) = b.Account_Key
INNER JOIN
(
SELECT business_lic, MAX(Lic_Exp_Date) AS Lic_Exp_Date
FROM Server.DB_B.dbo.TableB
GROUP BY business_lic, License_No
) c2 ON c2.business_lic = c1.business_lic AND c1.Lic_Exp_Date=c2.Lic_Exp_Date
WHERE DATEADD(dd,0, DATEDIFF(dd,0, c1.Lic_Exp_Date)) <> DATEADD(dd,0, DATEDIFF(dd,0, b.SEIN))
Here's what changed:
Turn the correlated subquery into a join. A join will be much faster, however the query optimizer might have been doing this for you already
Eliminate the need to call a number of per-row functions - should also help you match up with indexes better.
Use an integer comparison rather than string for your main join
Use date functions rather than convert to strings to remove the time portion, which should be much faster and in turn allow me to:
Use date comparisons rather than string comparisons in your join

Answer 2

The better way would be an Execute SQL task.