Import excel sheet to sql database VB Net client - vb.net

Question

I have some code I was hoping someone could look at for performance improvements. I have a spreadsheet that I need to import weekly. The sheet has 112 columns and about 35,000 rows. The code I have works, but it takes about 20 minutes to import the data. The excel column names do not match the database column names (I inherited this). Here is the code I am using. (I removed alot of the fields so it is easier to read)
'Connection String to Excel Workbook
Dim excelConnectionString As String = "Provider=Microsoft.ACE.OLEDB.12.0;Data Source=" & MyFile & ";Extended Properties=""Excel 12.0;HDR=Yes;"""
' Create Connection to Excel Workbook
Using connection As New System.Data.OleDb.OleDbConnection(excelConnectionString)
'List columns you need from the Excel file
Dim command As New System.Data.OleDb.OleDbCommand("Select * FROM [" & txtSheetName.Text & "$]", connection)
connection.Open()
' Create DbDataReader to Data Worksheet
Using dr As System.Data.OleDb.OleDbDataReader = command.ExecuteReader()
Dim strSql As String = ""
strSql = "INSERT INTO MyTestTable (" & _
"State, [Store Code], [Store Name], [Store Zone Code], [Store Zone Code Name], " & _
"WeekProd, YTDNew, " & _
"UpdatedBy, DateUpdated" & _
") VALUES (" & _
"#State, #StoreCode, #StoreName, #StoreZoneCode, #StoreZoneCodeName, " & _
"#WeekProd, #YTDNew, " & _
"#UpdatedBy, #DateUpdated" & _
Try
If dr.HasRows() Then
While dr.Read()
If Convert.ToString(dr.Item(0)) <> "" Then
Dim MyZone As String = Convert.ToString(dr.Item(1))
MyZone = StrConv(MyZone, vbProperCase)
Dim cmd As New SqlClient.SqlCommand
cmd.Connection = cn
cmd.CommandType = CommandType.Text
cmd.CommandText = strSql
cmd.Parameters.Add("#State", SqlDbType.VarChar).Value = ""
cmd.Parameters.Add("#StoreCode", SqlDbType.VarChar).Value = Convert.ToString(dr.Item(0))
cmd.Parameters.Add("#StoreName", SqlDbType.VarChar).Value = MyZone
cmd.Parameters.Add("#StoreZoneCode", SqlDbType.VarChar).Value = Convert.ToString(dr.Item(2))
cmd.Parameters.Add("#StoreZoneCodeName", SqlDbType.VarChar).Value = Convert.ToString(dr.Item(3))
cmd.Parameters.Add("#WeekProd", SqlDbType.Float).Value = Convert.ToDecimal(dr.Item(93))
cmd.Parameters.Add("#YTDNew", SqlDbType.Float).Value = Convert.ToDecimal(dr.Item(94))
cmd.Parameters.Add("#UpdatedBy", SqlDbType.VarChar).Value = MyUser.Substring(MyUser.Length - 4)
cmd.Parameters.Add("#DateUpdated", SqlDbType.Date).Value = Date.Today()
cmd.ExecuteScalar()
End If
End While
End If
Catch ex As Exception
lblMessage.Text = ex.Message
Exit Sub
Finally
cn.Close()
cn = Nothing
End Try
End Using
End Using

Answer 1

Won't promise this will set the world on fire performance-wise, but it might help.
My first thought would be to not create a new SqlCommand object and parameter set for every row in the Excel table. I would create it one time (which should save you the overhead of about 35K object instantiations given your data size), establish the parameter names, all outside the reader loop, and then call SetParameter to set the values for each column as each row in the Excel is traversed. That should trade you the overhead of about 35K * (number of real fields) parameter addition calls for setparameter calls to existing parameters. And I would also change the call type to ExecuteNonQuery rather than ExecuteScalar.
Now, ordinarily, the conventional wisdom would hold that you only open/close a connection to a database as you need it, and I think that is implied in this structure (open/close for each insert) but in this case, for a data update scenario like this, I think this would be a reasonable exception.