Question

I have a just one table mapped in a datacontext. Here’s the property and attribute on the column of interest:

[Column(Storage="_CustomerNumber", DbType="VarChar(25)")]
public string CustomerNumber
{

This column is, in fact, a varchar(25) and has an index.

I’ve got some simple code:

DataClasses1DataContext myDC = new DataClasses1DataContext();
myDC.Log = Console.Out;

List<string> myList = new List<string>() { "111", "222", "333" };
myDC.Customers
    .Where(c => myList.Contains(c.CustomerNumber))
    .ToList();

Which generates this SQL text:

SELECT [t0].[CustomerNumber], [t0].[CustomerName]
FROM [dbo].[Customers] AS [t0]
WHERE [t0].[CustomerNumber] IN (@p0, @p1, @p2)
-- @p0: Input NVarChar (Size = 3; Prec = 0; Scale = 0) [111]
-- @p1: Input NVarChar (Size = 3; Prec = 0; Scale = 0) [222]
-- @p2: Input NVarChar (Size = 3; Prec = 0; Scale = 0) [333]
-- Context: SqlProvider(Sql2005) Model: AttributedMetaModel Build: 3.5.21022.8

Notice that the paramaters are nvarchar!

When this query hits the database, it generates a horrible plan which involves converting the multi-million row index on CustomerNumber to nvarchar before seeking within it.

I’m not allowed to change the table, but I can change the query and the dbml. What can I do to get the data out without getting this index conversion?

Answer 1

Here’s the way I solve this problem now-a-days. This converts the parameters to the desired type and then runs the query. It generates the same sql as originally generated, just with different parameter types.

DbCommand myCommand = myDataContext.GetCommand(query);

foreach (DbParameter dbParameter in myCommand.Parameters)
{
  if (dbParameter.DbType == System.Data.DbType.String)
  {
    dbParameter.DbType = System.Data.DbType.AnsiString;
  }
}    

myDataContext.Connection.Open();

System.Data.Common.DbDataReader reader = myCommand.ExecuteReader();
List<RecordType> result = myDataContext.Translate<RecordType>(reader).ToList();

myDataContext.Connection.Close();