What is CDbl doing?

I had until recently been under the impression that the CDbl(x) operation in VB.NET was essentially a cast (i.e., the VB equivalent of (double)x in C#); but a recent discovery has revealed that this is not the case.

If I have this string:

Dim s As String = "12345.12345-"

And I do this:

Dim d As Double = CDbl(s)

d will be set to the value -12345.12345! Now, don't get me wrong, this is kind of convenient in my particular scenario; but I have to admit I'm confused as to why this works. In particular, I'm confused because:

• Double.Parse does not work with the above input.
• Double.TryParse does not work.
• Convert.ToDouble does not work.

How is CDbl so clever?

It uses Microsoft.VisualBasic.CompilerServices.Conversions.ToDouble(). That function contains a Select statement on the object's GetTypeCode() return value so it can use a custom converter based on the type of the argument. The string converter considers the possibility that the string might contain a currency value and does some processing on the string to deal with that. One allowed format for currency values is a trailing negative sign.

This is not particularly cheap. The quickest way to achieve the same conversion is:

Dim s As String = "12345.12345-"
Dim d As Double = Double.Parse(s, Globalization.NumberStyles.Any)

That's always been the behavior of CDbl() in Visual Basic 4/5/6 and is currently specific to VB.NET (it's inline, not part of the framework), so it's probably just been kept for people moving from earlier versions.

(Much like weirdness in pre-.NET Visual Basic due to features bought over from QBasic.)

If you go into Regional Options in Control Panel, there's a setting that let's you put the minus sign after, instead of before, numbers.

I'm not sure what system uses a minus sign after numbers, but it seems CDbl is programmed to accept both. Be liberal in what you accept, and all that.

Regional Options also has a setting for negative numbers where they are in brackets. Does that work? -- CDbl("(12345.12345)")