We have a value coming from an Oracle database, thru web svcs that has the data type of float (over zealous dba).  We need to display that value on a web page.  We originally had a generic conversion routine:

   public static decimal get_Decimal(object objIn)
        {
            decimal d = 0;
            if (objIn == null)
            { return d; }
            decimal.TryParse(objIn.ToString(), out d);         
            return d;
        }

This worked fine until we had a value of -15856966.  The TryParse failed and returned a 0. (The internal value was -1.58569E7, or something close to that.)  Of course, 0 would not satisfy the users!!!

We then tried:

if you are using oracle and have some values decimal whit  E over you will get and error

i think that you need save round values with oracle for dont get error 

Hi

I would recommend you to use System.Convert.ToDecimal(objIn) to convert the objIn to Decimal value. System.convert provides some intelligent roundings in numerical conversions, however the extra care taken by the System.Convert methods does come at a performance cost.

The alternative is to used Decimal.Round methods when you wanna do the decimal places rounding.

 I had tried the Convert.ToDecimal(objIn) and the String.Format methods, and they all worked find until the number got above 10,000,000, for example 15856966.  When it did, they returned a rounded number of 15856970.

What I finally got to work was:

decimal d = 0;
float f = (float)objIn;
double dblNbr = Convert.ToDouble( f );
d = Convert.ToDecimal( dblNbr );

This is pretty crazy. I am not sure if there is a bug in .Net or what.

You are working down in the murky, least significant bits of the float.  I think you will find that your float values are actually correct, because there is no way that just converting to a double can invent more bits of precision.  The problem is with the way the float is displayed, either in the debugger or via some form of Response.Write.  The default format string, I think, is "G", General which does not show all the stored digits of a float!  Try using the "Round-trip" format specifier when you display the float value.  Something like

string formattedNumber = string.Format("{0:R}", f);

where f is from line 2 of your code ( the cast objIn).  I suspect you will see that formattedNumber is correct.  You can get more information on the "R" specifier at http://msdn.microsoft.com/en-us/library/dwhawy9k.aspx#RFormatString

I failed to mention that this application is kind of a hodge-podge of platforms.  The db is Oracle, the middle tier is written in java, and the web ui is .Net.  I can see the value in the database and it is -15856966. However, when the TryParse executed it returned zero on the value that was passed thru the web svc (which also was verified to return the -1586966).  The standard ToString returned -1.58569E7.  I need to display the data in the UI, but when I do the display formatting String.Format( "{0:#,###,###,###.##;(#,###,###,###.##)}", it actually displays -15,856,970. So, not sure about the rounding issue.  However, the Format using the "R" does return the proper value.