We Keep Coding

Convert string to decimal but keep numeric portion

Is there a way to convert a string to decimal in C# but ignoring trailing "garbage"? i.e. like PHP's floatval() or C strtod() ?
e.g.
Convert string "2974.23abcdefs" to decimal 2974.23

As others have mentioned, there is no exact, like for like, replacement for what you can do in PHP and I think, for good reason. In the scenario of a web application, I'm not really sure that if I were accepting a decimal with garbage at the end, I'd actually want to consider that as valid data but this is just my opinion.
What you can do is define a regular expression that would capture the decimal and recognise that this is happening. I find this much safer and reliable.
Obviously, the regular expression can be improved but this is a simple example for you: -
var match = Regex.Match("2974.23abcdefs", "^([0-9]+\\.[0-9]+)(.+)?$");
if (match.Success)
{
// See the also the link below for using Decimal.TryParse
Console.WriteLine(Convert.ToDecimal(match.Groups[1].Value));
}
See https://msdn.microsoft.com/en-us/library/system.decimal.tryparse%28v=vs.110%29.aspx for my preferred way to convert to a decimal. This would ensure that you are coping with the output of the regular expression for how Decimal is comprised
For more information on regular expressions, see https://msdn.microsoft.com/en-us/library/system.text.regularexpressions.regex%28v=vs.110%29.aspx

This works but only takes care of digits and the current culture's decimal-separator:
string input = "2974.23abcdefs";
decimal d;
char decSep = NumberFormatInfo.CurrentInfo.NumberDecimalSeparator[0]; // there is no culture with a decimal-separator that has more than one letter so this isn't harmful
if(!string.IsNullOrEmpty(input) && Char.IsDigit(input[0]))
{
string number = string.Concat(input.TakeWhile(c => Char.IsDigit(c) || decSep == c));
bool validDecimal = decimal.TryParse(number, out d);
Console.WriteLine("Valid? {0} Parsed to: {1}", validDecimal, d);
}
Since we are using , as decimal separator here in germany i get a different result than people who use . as separator. You get 2974.23 and i get 2974.

As a first, second, third try, this should go:
static double Parse(string str, IFormatProvider provider = null)
{
if (str == string.Empty)
{
return 0;
}
if (provider == null)
{
provider = CultureInfo.CurrentCulture;
}
NumberFormatInfo nfi = NumberFormatInfo.GetInstance(provider);
// [ws][sign][integral-digits[,]]integral-digits[.[fractional-digits]][E[sign]exponential-digits][ws]
string ws = #"\s*";
string sign = #"(" + Regex.Escape(nfi.PositiveSign) + "|" + Regex.Escape(nfi.NegativeSign) + ")?";
string integralDigits1 = "([0-9](" + Regex.Escape(nfi.NumberGroupSeparator) + ")*)*";
string integralDigits2 = "[0-9]+";
string fractionalDigits = "(" + Regex.Escape(nfi.NumberDecimalSeparator) + "[0-9]*)?";
string exponentialDigits = "([Ee]" + sign + "[0-9]+)?";
var rx = new Regex(ws + sign + integralDigits1 + integralDigits2 + fractionalDigits + exponentialDigits);
string match = rx.Match(str).ToString();
if (match == string.Empty)
{
return 0;
}
return double.Parse(match, provider);
}
Note that the composed regex is very complex, because there are various "parts" in a full double that has been written to a string.
From MSDN:
[ws][sign][integral-digits[,]]integral-digits[.[fractional-digits]][E[sign]exponential-digits][ws]
Still some numbers will crash this function, if they are too much big. So passing new string('9', 1000) will make the double.Parse throw an exception.
Use it like:
double num = Parse(" +1,0.1234E+12abcdefgh", CultureInfo.InvariantCulture);
or
double num = Parse(" +1,,,,,0.1234E+12abcdefgh");
(if you don't need to configure the culture, will use the CultureInfo.CurrentCulture)

There are many ways to do so. I suggest using a Regex first, and then decimal.TryParse().
This is a regex that grabs a floating point number at the begin of the string, like -123.43 or just 1234.56 or 123456:
^([+-][0-9]+\.?[0-9]*).*$
Putting this into C# looks like this:
// Step 1: Getting some input
String input = "123.4533wefwe";
// Step 2: Get rid of the trail
Regex r = new Regex(#"^([+-][0-9]+\.?[0-9]*).*$", RegexOptions.IgnoreCase);
MatchCollection matches = r.Matches(input);
if (matches.Count > 0) {
Match match = matches[0];
GroupCollection groups = match.Groups;
// Step 3: create a real decimal from the string
decimal i;
NumberStyles style;
CultureInfo culture;
style = NumberStyles.Number;
culture = CultureInfo.CreateSpecificCulture("en-GB");
String matchedNumber = groups[1].Value;
if (decimal.TryParse(matchedNumber, style, culture, out i)) {
// Step 4: giving back the result:
Console.WriteLine("Parsed decimal: " + i);
}
}
The output of this is:
Parsed decimal: 123.4533
Remark: All this seems to become a bigger problem if you would like to parse real floating point number literals that include exponential notation. Then, severals stages of casting would be necessary.

Parse String to Decimal using any decimal seperator

I want to parse a string from an Text input to decimal. The value represents a currency value.
Currently i got this solution:
private Decimal CastToDecimal(string value)
{
Decimal result;
var valid = Decimal.TryParse(value, NumberStyles.Currency, null, out result);
return valid ? result : -1;
}
This works pretty well so far, except for possible culture-differences. I'm german and i expect most users to enter german-style puctuation. But it is possible that someone uses "." instead of "," and the conversion will fail.
"123,45€" => 123.45
"123.456,78€" => 123456.78
"123.45€" => 12345 <- I want the result to be 123.45 here
Is there a way to automatically detect the used culture for a decimal value? Such that it does not matter if you use german or english punctuation, you still get the same result?
Update:
Thanks to your help, i created a method which does what i want (i think).
private static Decimal CastToDecimal(string value)
{
Decimal resultDe;
Decimal resultEn;
var style = NumberStyles.AllowDecimalPoint | NumberStyles.AllowThousands;
var cultureDe = CultureInfo.CreateSpecificCulture("de-DE");
var cultureEn = CultureInfo.CreateSpecificCulture("en-GB");
var deValid = Decimal.TryParse(value, style, cultureDe, out resultDe);
var enValid = Decimal.TryParse(value, style, cultureEn, out resultEn);
var minVal = Math.Min(resultDe, resultEn);
var maxVal = Math.Max(resultDe, resultEn);
if (!deValid)
return resultEn;
if (!enValid)
return resultDe;
return BitConverter.GetBytes(decimal.GetBits(minVal)[3])[2] > 2 ? maxVal : minVal;
}
This code...
Console.WriteLine(CastToDecimal("123,45"));
Console.WriteLine(CastToDecimal("123.45"));
Console.WriteLine(CastToDecimal("123,450"));
Console.WriteLine(CastToDecimal("123.450"));
Console.WriteLine(CastToDecimal("123.123,45"));
Console.WriteLine(CastToDecimal("123,123.45"));
returns this:
123,45
123,45
123450
123450
123123,45
123123,45

the solution at http://msdn.microsoft.com/en-us/library/3s27fasw%28v=vs.110%29.aspx
which includes setting the NumberStyle may be helpful.
...
value = "1.345,978";
style = NumberStyles.AllowDecimalPoint | NumberStyles.AllowThousands;
culture = CultureInfo.CreateSpecificCulture("es-ES");
if (Double.TryParse(value, style, culture, out number))
Console.WriteLine("Converted '{0}' to {1}.", value, number);
else
Console.WriteLine("Unable to convert '{0}'.", value);
// Displays:
// Converted '1.345,978' to 1345.978.
value = "1 345,978";
if (Double.TryParse(value, style, culture, out number))
Console.WriteLine("Converted '{0}' to {1}.", value, number);
else
Console.WriteLine("Unable to convert '{0}'.", value);
...

I encountered the same problem some time ago. My solution was writing my own parser in Java. The algorithm first cleans up the string. Brief description follows:
Scan string from left to right
If char = '.' then dotFound=true ; lastSeparatorPosition = index ; dots++
If char = ',' then commaFound=true ; lastSeparatorPosition = index ; commas++
If dots == 0 && commas == 0 then its an integer => done
If dots > 0 && commas > 0 then the one at lastSeparatorPosition is the decimal separator. Remove the others from the string => done
/* only one separator type */ if ( dots + commas ) > 1 then remove them // because must be thousands separator => done
/* separator occurs once */ if numberOfDigits right of separator == 3 then you have to decide :-) either integer or decimal with 3 digits in fraction
7 is the only remaining problem like chiastic-security already stated. Here you can only decide taken the conceptual environment into account. All other cases are safe.
Have fun

This can't be done, simply because there are strings that are meaningful in two different cultures, but mean different things. For instance:
123.456
123,456
The first is a bit over 123 in the UK, but 123456 in Germany; the second is 123456 in the UK but a bit over 123 in France.

The only solution is to add validation on input and give user and example by the way if it is a webpage then find a way to get input according to user's culture. I suggests you to not to try to do what you are trying because there are some culture which contradict each others for example in currency;
US/Australia/Many others uses following format
45,999.95
where , is thousand separator and . is decimal separator
whereas in some European countries
45.999,95
means the same as above but thousands separator is . and , is used as decimal separator.
Now issue is there is no guarantee that user use both separator and your system may assume thousand separator as decimal and so on.
If you really don't want to bother user then make separate input fields for major and minor currencies.
So its better to not to go there. I believe this may help. Happy Coding :)
Update:
Same case is with date format e.g. in US format month comes first and then day whereas in Australia day comes first and then month now 02/01/2015 input will mean differently system can't tell the intention of user.

Strange output when converting string to double

I already searched for my problem but I wasn't successfully and that's the reason I'm here.
All I want to do is reading a string like "3.14" and convert it to double.
Enough said... here is my code:
using System;
namespace GlazerCalcApplication
{
class MainClass
{
public static void Main (string[] args)
{
string heightString;
double height;
heightString = Console.ReadLine();
height = Convert.ToDouble(heightString);
Console.WriteLine(height);
}
}
}
Output:
3.14
314
Press any key to continue...
Why is my double value not 3.14?
Instead of Convert.ToDouble() I also tried it with double.Parse() but I received the same behaviour. Reading strings like 3,14 is no problem.
Maybe I should also mention that I use MonoDevelop and a linux OS.
Thanks in advance.

Try specifying the culture as Invariant:
height = Convert.ToDouble(heightString,CultureInfo.InvariantCulture);
It seems the decimal seperator of your culture is comma instead of dot therefore dot is truncated after conversion.

Convert.ToDouble(string) uses Double.Parse(string, CultureInfo.CurrentCulture) method explicitly.
Here how it's implemented;
public static double ToDouble(String value) {
if (value == null)
return 0;
return Double.Parse(value, CultureInfo.CurrentCulture);
}
It is likely your CurrentCulture's NumberFormatInfo.NumberDecimalSeparator property is not . (dot). That's why you can't parse a string with . as a date seperator.
Example in LINQPad;
CultureInfo c = new CultureInfo("de-DE");
c.NumberFormat.NumberDecimalSeparator.Dump(); // Prints ,
As a solution, you can create a new reference of your CurrentCulture and assing it's NumberDecimalSeparator property to . like;
double height;
CultureInfo c = new CultureInfo("de-DE");
c.NumberFormat.NumberDecimalSeparator = ".";
height = Convert.ToDouble("3.14", c);

Judging by the result I take it you are in a culture zone where comma is the normal decimal separator.
Also, I take it that you want both dot and comma to be used for decimal separation.
If not, the below is not the proper solution.
The fastest solution for using both would be
height = Convert.ToDouble(heightString.Replace('.', ',');
This would mean that both dots and comma's are used as comma and thus parsed as a decimal separator.
If you only want to use a dot as separator, you can use invariantculture or a specific numberformatinfo. Invariant culture is already shown in the other posts. numberformat info example:
var nfi = new NumberFormatInfo { NumberDecimalSeparator = "." };
height = double.Parse(heightString,nfi);
For completeness, the example below shows both using numberformatinfo for setting the dot as decimal separator, as well as replacing comma with dots, so both characters are used for decimals
var nfi = new NumberFormatInfo { NumberDecimalSeparator = "." };
height = double.Parse(heightString.Replace(',', '.'),nfi);

Different .Net cultures (countries) have different decimal separators.
If you expect input values to be in some specific format - either use some particular culture or InvariantCulture. Also consider using double.Parse as it geve more flexibility on parsing the values than generic Convert.ToDouble.
var d = double.Parse(heightString, CultureInfo.InvariantCulture);
If you expect user to enter value in local format - your code is fine, but either your expectation of "local format" is wrong, or "current culture" set incorrectly.

How to replace a string with another

I'd like to make a program that isn't sensitive to the character , or . as input, in order to make some calculations without thinking about which symbol I should use to separate integer part from fractional part.
The only way i can imagine is to recognize the input string, check if there's a . and replace it with a , as my cultureType only accepts ,.
The problem is that I don't have any idea to make it, can someone help me?

if(inputString.Any(x => x == '.'))
{
inputString = inputString.Replace('.', ',');
}

I'm guessing (it's not in your question) that you're trying to parse a decimal, and the number may have a decimal separator and no thousand separator, and it might be a , and it might be a . A possible heuristic to guess which one it is, is to check the last occurrence of either.
That could be done by
string input = "123.456.496,58"; //or 123,456,789.58
int lastComma = input.LastIndexOf(",");
int lastPeriod = input.LastIndexOf('.');
NumberFormatInfo format = new NumberFormatInfo();
if (lastComma > lastPeriod) {
format.NumberDecimalSeparator = ',';
format.NumberGroupSeparator = '.';
} else {
format.NumberDecimalSeparator = '.';
format.NumberGroupSeparator = ',';
}
double parsed = Double.Parse(intput, format);
this fails when there is a group separator after the decimal separator, and other heurisics could be employed to change it (i.e. if there is only one non-numeric character, that one is the decimal separator. If there are multiple non-numeric special characters, and one of them only occurs once, that is the decimal separator)
In the end parsing a string with an unknown format will require some guesswork.
If you're not trying to parse numbers, but are doing something different entirely, disregard this answer, and pick one of the others.

if (yourString.Contains(".")) { youString = yourString.Replace(".", ","); }

Format a double value like currency but without the currency sign (C#)

I feed a textbox a string value showing me a balance that need to be formatted like this:
###,###,###,##0.00
I could use the value.ToString("c"), but this would put the currency sign in front of it.
Any idea how I would manipulate the string before feeding the textbox to achieve the above formatting?
I tried this, without success:
String.Format("###,###,###,##0.00", currentBalance);
Many Thanks,

If the currency formatting gives you exactly what you want, clone a NumberFormatInfo with and set the CurrencySymbol property to "". You should check that it handles negative numbers in the way that you want as well, of course.
For example:
using System;
using System.Globalization;
class Test
{
static void Main()
{
NumberFormatInfo nfi = CultureInfo.CurrentCulture.NumberFormat;
nfi = (NumberFormatInfo) nfi.Clone();
Console.WriteLine(string.Format(nfi, "{0:c}", 123.45m));
nfi.CurrencySymbol = "";
Console.WriteLine(string.Format(nfi, "{0:c}", 123.45m));
}
}
The other option is to use a custom numeric format string of course - it depends whether you really want to mirror exactly how a currency would look, just without the symbol, or control the exact positioning of digits.

string forDisplay = currentBalance.ToString("N2");

Have you tried:
currentBalance.ToString("#,##0.00");
This is the long-hand equivalent of:
currentBalance.ToString("N2");

string result=string.Format("{0:N2}", value); //For result like ### ### ##.##

You can do this with the group separator and the section separator, like this:
currentBalance.ToString("#,0.00;(#,0.00)");
This does not account for culture variances like the answer from #JonSkeet would, but this does mimic decimal place, rounding, thousands separation, and negative number handling that en-US culture currency format produces using a single custom format string.
.NET Fiddle Demo

var result = currentBalance.ToString("C").Replace(System.Globalization.CultureInfo.CurrentCulture.NumberFormat.CurrencySymbol, "");

CultureInfo cultureInfo = new CultureInfo("en-US");
cultureInfo.NumberFormat.CurrencySymbol = "Rs.";
Thread.CurrentThread.CurrentCulture = cultureInfo;
decimal devimalValue = 3.45M;
this.Text = devimalValue.ToString("C2"); //Rs.3.45

This may be overkill, but it rounds, formats...
#helper TwoDecimalPlaces(decimal? val)
{
decimal x = 0;
decimal y = 0;
string clas = "text-danger";
if (val.HasValue)
{
x = (decimal)val;
if (val > 0)
{
clas = "";
}
}
y = System.Math.Round(x, 2);
IFormatProvider formatProvider = new System.Globalization.CultureInfo(string.Empty);
<span class="#clas">#string.Format("{0:N2}", y)</span>
}

This simple solution works for me with US currency.
If not needing international currency support use this and replace the $ with the currency symbol(s) to be removed:
// for USD
string result = currentBalance.ToString("C").Replace("$", "")
or
// for EUR
string result = currentBalance.ToString("C").Replace("€", "")