Need fast ListOf(type comparison

[!NoItAll]

I have a ListOf(Type that I've created that will often have well over 20-thousand entries.
The type I am placing into the list looks like this:

Structure MyType
   Dim FileName as String
   Dim FileDate as String
   Dim FileSize as String
   Dim NewPath as String
End Structure
Dim MyList as ListOf(MyType)
 
 

Before I add anything to the list I need to check and make sure it's not already in there:

If  MyList.Contains(sNewEntry) = False then
   MyList.Add(sNewEntry)
End if
 

...but with anything above about 7-thousand entries it takes about 1/4 to 1/2 second for each comparison. With more than 15-thousand it takes nearly a second. If I am adding a total of 20-thousand records I could miss Christmas!

I *thought* it would be faster to try

If MyList.BinarySearch(sNewEntry) < 0 then
   MyList.Add(sNewEntry)
End if
 

But this throws an exception ("failed to compare") - so either this is not what I want or I am just doing it wrong.

I also thought I might throw a hash into the structure and then force the CONTAINS or BINARYSEARCH method to only compare hashes - but I just can't figure out how to do that... Also - just how unique is the default hash?

Any suggestions would be very much appreciated!

If I stumble across a better solution I will be sure to post it here - I loath threads without answers!

Thanks all!

[tlhintoq]

So each list element is a struct of 4 items.
So each test of the compare is actually having to check if all 4 items in the list element are the same as your comparitor.
Meaning 5,000 element test is 20,000 {internal} comparisons.

If the struct were a class you could access the members individually.
MyType.FileName for example
You could the loop through testing any one element to see if its the same.
Now you would be doing 5,000 comparisons instead of 20,000

[!NoItAll]

Here's what I've done - it's not big trick, but it has improved performance dramatically. As tlhintoq pointed out there were way too many compares going on.
I added an additional element to the structure

Structure MyType
   Dim FileName as String
   Dim FileDate as String
   Dim FileSize as String
   Dim NewPath as String
   Dim FileNameHash as Integer
End Structure
Dim MyList as ListOf(MyType)
 

Then I wrote my own compare routine

Friend Function CompareHash(ByRef LongList as List(of MyType), NewItem as MyType) as Boolean
 
For Each item as MyType in LongList
    If item.FileNameHash = NewItem.FileNameHash then
           If Item.FileName = NewItem.FileName then
                 Return True
            End if
     End if
Next
Return False
 

This way I am only looking at the hash (an integer) so the checks will be very fast.
Notice that I check the filename even after there is a hash match. This is because the hash is not guarenteed unique. If I have two strings they will definately generate the exact same hash. But it is possible for two different strings to generate identical hashes - so we do a second compare after matching hashes are found to be sure. This is still hundreds of times faster than doing all string matches.
For my routine a list of 22-thousand items went from 5 minutes to 18 seconds.