Data Adapter Vs Sql Command

Question

Which one would be better in executing an insert statement for ms-sql database:

Sql DataAdapter or SQL Command Object?

Which of them would be better, while inserting only one row and while inserting multiple rows?

A simple example of code usage:

SQL Command

string query = "insert into Table1(col1,col2,col3) values (@value1,@value2,@value3)";
int i;

SqlCommand cmd = new SqlCommand(query, connection);
// add parameters...
cmd.Parameters.Add("@value1",SqlDbType.VarChar).Value=txtBox1.Text;
cmd.Parameters.Add("@value2",SqlDbType.VarChar).Value=txtBox2.Text;
cmd.Parameters.Add("@value3",SqlDbType.VarChar).Value=txtBox3.Text;
cmd.con.open();
i = cmd.ExecuteNonQuery();
cmd.con.close();

SQL Data Adapter

DataRow dr = dsTab.Tables["Table1"].NewRow();
DataSet dsTab = new DataSet("Table1");
SqlDataAdapter adp = new SqlDataAdapter("Select * from Table1", connection);
adp.Fill(dsTab, "Table1");
dr["col1"] = txtBox1.Text;
dr["col2"] = txtBox5.Text;    
dr["col3"] = "text";

dsTab.Tables["Table1"].Rows.Add(dr);

SqlCommandBuilder projectBuilder = new SqlCommandBuilder(adp);
DataSet newSet = dsTab.GetChanges(DataRowState.Added);
adp.Update(newSet, "Table1");

Answer 1

Updating a data source is much easier using DataAdapters. It's easier to make changes since you just have to modify the DataSet and call Update.

There is probably no (or very little) difference in the performance between using DataAdapters vs Commands. DataAdapters internally use Connection and Command objects and execute the Commands to perform the actions (such as Fill and Update) that you tell them to do, so it's pretty much the same as using only Command objects.

Answer 2

I would use LinqToSql with a DataSet for single insert and most Database CRUD requests. It is type safe, relatively fast for non compilcated queries such as the one above.

If you have many rows to insert (1000+) and you are using SQL Server 2008 I would use SqlBulkCopy. You can use your DataSet and input into a stored procedure and merge into your destination

For complicated queries I recommend using dapper in conjunction with stored procedures.