如何使.NET Hashtable像Java Hashtable一样工作

Question

When I add some items into a Java Hashtable , their order is different to that of a .NET Hashtable . Is there any way I can make sure the .NET Hashtable will have the same order as a Java Hashtable ?

---

I'm trying to port some Java code to C#. The Java code uses a Hashtable to keep track of some data. When I check to see the order the data is retrieved when I iterate through either the Java Hashtable or the .NET Hashtable (via an Enumerator ), each one consistently has the same data, in the same order ... but each code based has a different order.

Is there any way I can make it so that the .NET Hashtable data is in the same order as the Java Hashtable ?

I understand that Hashtable do not handling ordering - so I feel like there's nothing that can be done. I also cannot change the datatype in the Java code from .. say .. a Hashtable to something else.

Here's some same data to illustrate my situation.

Data, added sequential for either code base :-

1. num | someData
2. pagenum | someData
3. x | someData
4. top | someData

Java Code:

private Hashtable identifiers = new Hashtable();
...
identifiers.put(symbol, identifier);

Java output via iteration over an enumerator:

.NET code:

private Hashtable Identifiers = new Hashtable();
...
Identifiers.Add(symbol, identifier);

.NET output via iteration over an enumerator.

Any ideas or suggestions?

Answer 1

Unfortunately, short of changing the type of the collection there's really no way around this. Hashtables don't guarantee order (as you mentioned in your post).

It all boils down to the following:

If order matters, a Hashtable is the wrong data structure.

Answer 2

You cannot change this order short of extracting the keys and sorting them in some way (or using an ordered dictionary). The iteration orders of either of the hashtable implementations you are using are not specified, and therefore relying on a specific order means your code is broken. The iteration order in Java is not even guaranteed to remain the same in future versions of the JDK.

Answer 3

The order in the generic hashtable is unpredictable. That's what you are experiencing. If you are trying to run comparison tests and if you need specific order, you have to use ordered maps on BOTH ends. Otherwise there isn't hardly a way to predict the output.

Answer 4

I agree with other answers, but still, Hashtable is deprecated. You should always use a generic Dictionary class. There is also a SortedDictionary, class name says all. you can reverse or sort it's content, take a look here Reverse Sorted Dictionary in .NET

Answer 5

You could use a LinkedHashMap in Java and Recursion Software makes one available for C# as per this post .

I have not tried this solution.

Answer 6

It's a terrible solution, but according to your description you have no other option.

You can take sources of Java Hastable class and port them to C#. Strip down everything unnecessary, like iterators and keySet , equals , and you will have pretty small class (few dozen lines I'd expect). Converting it to C# should be trivial.

Naturally, you shouldn't use this new class for anything other than Java compatibility. And filing a bug with original application vendor might be nice too :)

Answer 7

You could try to implement (in .NET) your own IHashCodeProvider and IComparer and pass those to the Hashtable constructor. From the example you posted it appears that the entries are sorted on the key ascending (.NET) and descending (Java) order so you'd need a comparer that inverts the order.