I recently inherited a C# Web app that creates a new connection for every query like so:
public class QueryForm { public bool getStudents() { SqlConnection conn = new SqlConnection(ConfigurationManager.ConnectionStrings["conn"].ConnectionString); conn.Open(); //commands conn.Close(); } protected void getProfessors() { SqlConnection conn = new SqlConnection(ConfigurationManager.ConnectionStrings["conn"].ConnectionString); conn.Open(); //Commands conn.Close(); } protected void getProfessors() { SqlConnection conn = new SqlConnection(ConfigurationManager.ConnectionStrings["conn"].ConnectionString); conn.Open(); //Commands conn.Close(); } }
I understand that this is typically the best way to do it, but is it acceptable or “best practice” to have the constructor create the connection object, and then have each method/Query open and then close that connection like so:
public class QueryForm { SqlConnection conn; public QueryForm() { conn = new SqlConnection(ConfigurationManager.ConnectionStrings["conn"].ConnectionString); } public bool getStudents() { conn.Open(); //commands conn.Close(); } protected void getProfessors() { conn.Open(); //Commands conn.Close(); } protected void getCourses() { conn.Open(); //Commands conn.Close(); } }
I prefer the second way as it does not make multiple connection objects. However, the first way would also be preferable if I were to make the methods and class static.
Advertisement
Answer
Either one of these are acceptable. A SqlConnection
uses a connection-pool so it shouldn’t affect performance much. Having multiple SqlConnection
objects isn’t going to hurt anything. This comes down to a preference.
I would suggest encapsulating the commands in a using
statement if you keep the connections inside the methods, such as:
using (SqlConnection conn = new SqlConnection(...)) { conn.Open(); //commands conn.Close(); }
This ensures proper disposal of the connection.