|
AKAIK, you can only write code against v1.1 of the framework in VS2003.
You need to either upgrade to VS2005, or use a 3rd party IDE like SharpDevelop.
Paul Marfleet
"No, his mind is not for rent
To any God or government"
Tom Sawyer - Rush
|
|
|
|
|
Thanks. So what about .NET 3.0 ? If I have VS2005 installed, can I use .NET 3.0 classes ? I guess we can
|
|
|
|
|
Yes, you can use dot net 3 in VS2005, but not in VS2003.
"You're very clever, young man, very clever," said the old lady. "But it's turtles all the way down!"
|
|
|
|
|
This link[^] explains how to update VS2005 to access .NET 3.0 functionality.
I believe that in VS2008, it will be possible to build your code against a specific version of the framework.
Paul Marfleet
"No, his mind is not for rent
To any God or government"
Tom Sawyer - Rush
|
|
|
|
|
Thanks for the information paul.
|
|
|
|
|
pmarfleet wrote: I believe that in VS2008, it will be possible to build your code against a specific version of the framework.
Your right but it's limited to 2.0, 3.0 and 3.5, you wont be able to do 1.0 or 1.1.
|
|
|
|
|
You can compile for .NET 1.0 and 1.1 in both VS2005 and VS2008, you just have to write the appropriate MSBuild targets (or download them, e.g. from here[^]).
|
|
|
|
|
I need to represent some data in memory which needs the performance of a
Dictionary lookup, but the key for these data consists of multible values,
e.g.:
Data has a structure like this:
ProductId (key) InventId (key) Property1 (value) Property2 (value)
1234 1 [somevalue] [somevalue]
1234 2 [somevalue] [somevalue]
1237 1 [somevalue] [somevalue]
The unique key in this case is a combination of 'ProductId' and 'InventId'
My first idea was to startout having the key a new Generic class e.g.
Key<T1,T2> for use with Dictionary<Key<T1,T2>, ProductRelation> etc..
- But this solution is not usable for more than two values for the key ....
What is the "best" Practice for this scenario? - if you want max flexibility and max
performance ??
Regards, Rune
|
|
|
|
|
Hello,
Why not use a non generic class or struct?
All the best,
Martin
|
|
|
|
|
a nongeneric collection would be a lot slower.
I could use a object[] as the key, but that would give a heavy performance penalty when boxing/unboxing values ... that's why generics is so good in this case.
|
|
|
|
|
Create a wrapper object that overrides the GetHashCode method.
Actually, when it is numeric I just make my hash use a string as the first key and look up "id_id"
The other option is to have a hash containing a hash. It would require two look ups but work reasonably well.
Dictionary<productid, dictionary<inventoryid,="" myobject="">>;
Need a C# Consultant? I'm available.
Happiness in intelligent people is the rarest thing I know. -- Ernest Hemingway
|
|
|
|
|
This performs quite well, but it still does not solve having n-number of values as a key, only two this time ...
<br />
void Main()<br />
{<br />
Dictionary<Key<int, string>, MyClass> dictionary = new Dictionary<Key<int, string>, MyClass>();<br />
<br />
MyClass obj = new MyClass();<br />
dictionary.Add(obj.GetKey(), obj);<br />
}<br />
<br />
<br />
<br />
<br />
public struct Key<T1, T2> : IEquatable<Key<T1, T2>><br />
{<br />
public Key(T1 keyValue1, T2 keyValue2)<br />
{<br />
_keyValue1 = keyValue1;<br />
_keyValue2 = keyValue2;<br />
}<br />
<br />
private T1 _keyValue1;<br />
private T2 _keyValue2;<br />
<br />
public override int GetHashCode()<br />
{<br />
return _keyValue1.GetHashCode() ^ _keyValue2.GetHashCode();<br />
}<br />
<br />
public override bool Equals(object other)<br />
{<br />
return this.GetHashCode().Equals(other.GetHashCode());<br />
}<br />
<br />
}<br />
<br />
<br />
<br />
public sealed class MyClass<br />
{<br />
<br />
private int _Key1;<br />
private string _Key2;<br />
private string _Value1;<br />
<br />
public int Key1<br />
{<br />
get { return _Key1; }<br />
set { _Key1 = value; }<br />
}<br />
<br />
public string Key2<br />
{<br />
get { return _Key2; }<br />
set { _Key2 = value; }<br />
}<br />
<br />
public string Value1<br />
{<br />
get { return _Value1; }<br />
set { _Value1 = value; }<br />
}<br />
<br />
public Key<int, string> GetKey()<br />
{<br />
return new Key<int, string>(this.Key1, this.Key2);<br />
}<br />
<br />
}<br />
<br />
|
|
|
|
|
How about a Dictionary of Dictionaries?
I don't know how variable you can get though.
Gives me something to think about tonight...
|
|
|
|
|
Under the right circumstances it would probably perform very well, but for more than two values you are screwed...
|
|
|
|
|
|
I like a good programming exercise. What I've been concentrating on is the "max flexibility" part, but with a lack of type safety  I don't see how both can be achieved.
This is just a little test program that uses the class I'm working on.
namespace Test
{
class Test
{
[System.STAThreadAttribute]
static void
Main
(
string[] args
)
{
PIEBALD.Types.Hierarchary store = new PIEBALD.Types.Hierarchary
(
typeof(int)
,
typeof(int)
,
typeof(string)
) ;
store [ 1234 , 1 , "Property1" ] = "1234 1 1" ;
store [ 1234 , 1 , "Property2" ] = "1234 1 2" ;
store [ 1234 , 2 , "Property1" ] = "1234 2 1" ;
store [ 1234 , 2 , "Property2" ] = "1234 2 2" ;
store [ 1237 , 1 , "Property1" ] = "1237 1 1" ;
store [ 1237 , 1 , "Property2" ] = "1237 1 2" ;
System.Console.WriteLine
( "ProductId (key) InventId (key) Property1 (value) Property2 (value)" ) ;
foreach ( object k1 in store.Keys() )
{
foreach ( object k2 in store.Keys ( k1 ) )
{
System.Console.WriteLine
(
"{0,-17} {1,-16} {2,-20} {3,-20}"
,
k1
,
k2
,
store [ k1 , k2 , "Property1" ]
,
store [ k1 , k2 , "Property2" ]
) ;
}
}
return ;
}
}
}
|
|
|
|
|
Clever, but you don't avoid the unfortunate cast to and from object without the use of generics...
I've been working on a generic KeyedList, a kind of "brother" to the KeyedCollection, but for use with multible keys.
It's used like this
<code>
KeyedList<string, int, MyObject> keyedlist = new KeyedList<string, int, MyObject>();
keyedlist.Add(new MyObject("kvalue1",123, "actual content"));
keyedlist.Add(new MyObject("kvalue2",123, "actual content"));
// etc...
string key1 = "kvalue1";
int key2 = 123;
if(keyedlist.Contains(key1, key2);
{
MyObject obj = keyedlist[key1, key2];
// etc..
}
</code>
but I had make a new class for each number of keytypes for a coollection...
For 1, 2 and 3 keys the outcome was:
<code>
class KeyedList<TKey1, TObject>
{
}
class KeyedList<TKey1, TKey2, TObject>
{
}
class KeyedList<TKey1, TKey2, TKey3, TObject>
{
}
// etc...
</code>
I was hoping to avoid that ...
My dreams was that you could declare n-number of TKey types in a generic class....
Something like this:
<code>
class KeyedList<TKey[], TObject>
{
void Somemethod()
{
foreach(TKey keytype in TKey[])
{
// etc...
}
}
}
// so you could use it like this
KeyedList<int, int, Guid, MyObject> list = new ...
// or
KeyedList<string, MyObject> list = new ...
// with out having to make new class declare for each ....
</code>
Maybe in the next C# version...
|
|
|
|
|
I don't think the key types matter as much as the value type, generally they go in but don't come out.
Because I was aiming for max flexibility I didn't provide the ability to specify the value type, but it could be done.
|
|
|
|
|
I added the ability to specify the value type and the key type when desired.
namespace Test
{
class Test
{
[System.STAThreadAttribute]
static void
Main
(
string[] args
)
{
try
{
PIEBALD.Types.Hierarchary<string> store =
new PIEBALD.Types.Hierarchary<string>
(
typeof(int)
,
typeof(int)
,
typeof(string)
) ;
store [ 1234 , 1 , "Property1" ] = "1234 1 1" ;
store [ 1234 , 1 , "Property2" ] = "1234 1 2" ;
store [ 1234 , 2 , "Property1" ] = "1234 2 1" ;
store [ 1234 , 2 , "Property2" ] = "1234 2 2" ;
store [ 1237 , 1 , "Property1" ] = "1237 1 1" ;
store [ 1237 , 1 , "Property2" ] = "1237 1 2" ;
System.Console.WriteLine ( "ProductId (key) InventId (key) Property1 (value) Property2 (value)" ) ;
foreach ( int k1 in store.Keys<int>() )
{
foreach ( int k2 in store.Keys<int> ( k1 ) )
{
System.Console.WriteLine
(
"{0,-17} {1,-16} {2,-20} {3,-20}"
,
k1
,
k2
,
store [ k1 , k2 , "Property1" ]
,
store [ k1 , k2 , "Property2" ]
) ;
}
}
}
catch ( System.Exception err )
{
System.Console.WriteLine ( err.Message ) ;
}
return ;
}
}
}
|
|
|
|
|
I have a method which is passed an Exception.
I'd like to find out which Exception it is (ie, was it ArgumentNullException, was it a COM Exception, etc).
How do I address this? I don't have control of the original Try/Catch statement in this method, it just receives an Exception which is passed, not thrown.
|
|
|
|
|
if(ex is ArgumentNullException)
{
MessageBox.Show("Thau shall not nullify");
}
fafafa, ringakta icke sådant som bringa ack så naggande högönsklig välmåga å baronens ära.
|
|
|
|
|
Thanks, that was exactly what I was looking for - what comparator to use. I was trying "==" and getting nowhere. Sounds like "is" is the way to go.
I was also just casting from the general exception to the specific exception and seeing what happened. I was not getting much info there  .
|
|
|
|
|
How can I access the specific fields and properties?
public string GetErrorCode(Exception ex)<br />
{<br />
if (ex is COMException)<br />
{<br />
return ex.ErrorCode.ToString();<br />
}<br />
else<br />
{<br />
return "";<br />
}<br />
}
This doesn't work, the compiler complains that System.Exception doesn't contain a definition for ErrorCode.
I tried casting ex using
ex = (COMException)ex
but it fails the same way.
|
|
|
|
|
Hello,
You have to cast it to the Type!
if (ex is COMException)
{
COMException comEx = ex as COMException
return comEx.ErrorCode.ToString();
}
or
COMException comEx = ex as COMException
if(comEx !=null)
{
return comEx.ErrorCode.ToString();
}
All the best,
Martin
|
|
|
|
|
Interesting, I used a direct cast (as in my example posted further down) instead of "as", and it appears to work fine. Any caveats there I should know about?
I'm not that familiar with "is" and "as", clearly. Perhaps I should be...
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
