c# - What are the best practices to parse data into objects in data access layer? -

i adding new features old asp.net application uses n-tier architecture. basic example can given as

object

public class test {     public int id{get;set;}     public int name{get;set;} } 

data access layer

public static list<test> gettests() {     list<test> list = new list<test>();     try     {         //codes         sqldatareader dr  = com.executereader();         while(dr.read())             list.add(filltestrecord(dr))         //codes     }     catch{}     return list; }  private static test filltestrecord(idatarecord dr) {     test test = new test();     try{test.id = convert.toint32(dr["id"]);}     catch{}     try{test.name = convert.toint32(dr["name"]);}     catch{}     return test; } 

the development requires me add new fields object classes , re-usability use 1 fill*record method each type of object. method can called many other dal methods idatarecord might not contain columns of object. hence put try-catch block every property separately. ensures available columns in idatarecord parsed properly.

my question is there better way of doing it? , best practices in type of architecture?

update

after reading comment/answer of david l , anup have tried way using extension method. method follows

public static bool trygetordinal(this idatarecord dr, string column, out int ordinal) {     try     {         ordinal = dr.getordinal(column);     }     catch(exception ex)     {         ordinal = -1; //just setting value getordinal doesn't return         return false;     }     return true; } 

so filltestrecord method be

private static test filltestrecord(idatarecord dr) {     test test = new test();      int ordinal = default(int);      if(dr.trygetordinal("id",out ordinal))         test.id = convert.toint32(dr.getvalue(ordinal));     if(dr.trygetordinal("name",out ordinal))         test.name = convert.tostring(dr.getvalue(ordinal));      return test; } 

any suggestion on highly appreciated.

update 03-02-2016

during debugging found try-catch takes big toll on performance if getordinal throws error when supplied column name not found in datarecord. wrote new method gets column names in datareader , replaced getordinal array.indexof.

public static bool trygetordinal(this idatarecord dr, string[] columnnames, string column, out int ordinal) {     ordinal = array.indexof(columnnames, column);     return ordinal >= 0; } 

and filltestrecord becomes -

private static test filltestrecord(idatarecord dr, string[] columnnames) {     test test = new test();      int ordinal = default(int);      if(dr.trygetordinal(columnnames, "id",out ordinal))         test.id = convert.toint32(dr.getvalue(ordinal));     if(dr.trygetordinal(columnnames, "name",out ordinal))         test.name = convert.tostring(dr.getvalue(ordinal));      return test; } 

column names passed fill method -

using (var dr = com.executereader()) {     string[] colnames = dr.getcolumnnames();     while (dr.read())         list.add(filltestrecord(dr, colnames)); } 

'getcolumnnames' new extension method -

public static string[] getcolumnnames(this idatareader dr) {     string[] columnnames = new string[dr.fieldcount];     (int = 0; < dr.fieldcount; i++)     {         columnnames[i] = dr.getname(i).tolower();     }     return columnnames; } 

seems me in right direction.
long parsing being done in centralized location re-used upper level classes, looks solution.

the thing change replacing try-catch statements checking if data exists in columns. surely there way tell (column not exist? db-null value?) implement using similar tryparse methods.

    private static test filltestrecord(idatarecord dr)     {         test test = new test();         int tempid;         if (tryparsedatarow<int>(dr, "id", out tempid))         {             test.id = tempid;         }         return test;     }      private static bool tryparsedatarow<t>(idatarecord record, string column, out t value)     {         value = default(t);         bool success = true;          if (record == null)         {             //nothing can null object             success = false;         }         else if (!record.hascolumn(column)) //not sure if throw exeption or return null. can check in project         {             success = false;         }         else if (record[column] != typeof(t))         {             //object of unexpected type             success = false;         }                    else         {             //cast value output parameter             value = (t)record[column];         }          return success;     } 

and of course have implement hascolumn method (implemented here extension):

    /// <summary>     /// determines whether specified record has column.     /// </summary>     /// <param name="record">the record.</param>     /// <param name="columnname">name of column.</param>     /// <returns>true if column exist, false otherwise</returns>     public static bool hascolumn(this idatarecord record, string columnname)     {         (int = 0; < record.fieldcount; i++)         {             if (record.getname(i).equals(columnname, stringcomparison.invariantcultureignorecase))                 return true;         }         return false;     }