DbDataRecord.cs source code in C# .NET

Source code for the .NET framework in C#

                        
Code:

                         / 4.0 / 4.0 / untmp / DEVDIV_TFS / Dev10 / Releases / RTMRel / ndp / fx / src / Data / System / Data / Odbc / DbDataRecord.cs / 1305376 / DbDataRecord.cs
                        

                        
                            //------------------------------------------------------------------------------ 
// 
//      Copyright (c) Microsoft Corporation.  All rights reserved.
// 
// [....] 
// [....]
//----------------------------------------------------------------------------- 
 
using System;
using System.ComponentModel;            //Component 
using System.Data;
using System.Runtime.InteropServices;   //Marshal
using System.Reflection;                //Missing
 
namespace System.Data.Odbc {
    sealed internal class DbSchemaInfo { 
 
        internal DbSchemaInfo() {
        } 

        internal string _name;
        internal string _typename;
        internal Type _type; 
        internal ODBC32.SQL_TYPE? _dbtype;
        internal object _scale; 
        internal object _precision; 

        // extension to allow BindCol 
        //
        internal int _columnlength;          //
        internal int _valueOffset;           // offset to the data in the row buffer
        internal int _lengthOffset;          // offset to the length in the row buffer 
        internal ODBC32.SQL_C _sqlctype;         // need this to bind the value
        internal ODBC32.SQL_TYPE _sql_type;      // need that to properly marshal the value 
    } 

    ///////////////////////////////////////////////////////////////////////////// 
    // Cache
    //
    //  This is a on-demand cache, only caching what the user requests.
    //  The reational is that for ForwardOnly access (the default and LCD of drivers) 
    //  we cannot obtain the data more than once, and even GetData(0) (to determine is-null)
    //  still obtains data for fixed lenght types. 
 
    //  So simple code like:
    //      if(!rReader.IsDBNull(i)) 
    //          rReader.GetInt32(i)
    //
    //  Would fail, unless we cache on the IsDBNull call, and return the cached
    //  item for GetInt32.  This actually improves perf anyway, (even if the driver could 
    //  support it), since we are not making a seperate interop call...
 
    //  We do not cache all columns, so reading out of order is still not 
    //
    ///////////////////////////////////////////////////////////////////////////// 
    sealed internal class DbCache {
        //Data

        private bool[] _isBadValue; 
        private DbSchemaInfo[] _schema;
        private object[] _values; 
        private OdbcDataReader _record; 
        internal int _count;
        internal bool _randomaccess = true; 

        //Constructor
        internal DbCache(OdbcDataReader record, int count) {
            _count = count; 
            _record = record;
            _randomaccess = (!record.IsBehavior(CommandBehavior.SequentialAccess)); 
            _values = new object[count]; 
            _isBadValue = new bool[count];
        } 

        //Accessor
        internal object this[int i] {
            get { 
                if(_isBadValue[i]) {
                    OverflowException innerException = (OverflowException)Values[i]; 
                    throw new OverflowException(innerException.Message, innerException); 
                }
                return Values[i]; 
            }
            set {
                Values[i] = value;
                _isBadValue[i] = false; 
            }
        } 
 
        internal int Count {
            get { 
                return _count;
            }
        }
 
        internal void InvalidateValue(int i) {
            _isBadValue[i] = true; 
        } 

        internal object[] Values { 
            get {
                return _values;
            }
        } 

        internal object AccessIndex(int i) { 
            //Note: We could put this directly in this[i], instead of having an explicit overload. 
            //However that means that EVERY access into the cache takes the hit of checking, so
            //something as simple as the following code would take two hits.  It's nice not to 
            //have to take the hit when you know what your doing.
            //
            //  if(cache[i] == null)
            //      .... 
            //  return cache[i];
 
            object[] values = this.Values; 
            if(_randomaccess) {
                //Random 
                //Means that the user can ask for the values int any order (ie: out of order).
                //  In order to acheive this on a forward only stream, we need to actually
                //  retreive all the value in between so they can go back to values they've skipped
                for(int c = 0; c < i; c++) { 
                    if(values[c] == null) {
                        values[c] = _record.GetValue(c); 
                    } 
                }
            } 
            return values[i];
        }

        internal DbSchemaInfo GetSchema(int i) { 
            if(_schema == null) {
                _schema = new DbSchemaInfo[Count]; 
            } 
            if(_schema[i] == null) {
                _schema[i] = new DbSchemaInfo(); 
            }
            return _schema[i];
        }
 
        internal void FlushValues() {
            //Set all objects to null (to explcitly release them) 
            //Note: SchemaInfo remains the same for all rows - no need to reget those... 
            int count = _values.Length;
            for(int i = 0; i < count; ++i) { 
                _values[i] = null;
            }
        }
    } 
}

// File provided for Reference Use Only by Microsoft Corporation (c) 2007.

Link Menu

Network programming in C#, Network Programming in VB.NET, Network Programming in .NET
This book is available now!
Buy at Amazon US or
Buy at Amazon UK