แก้ไข

แชร์ผ่าน


Parsing numeric strings in .NET

All numeric types have two static parsing methods, Parse and TryParse, that you can use to convert the string representation of a number into a numeric type. These methods enable you to parse strings that were produced by using the format strings documented in Standard Numeric Format Strings and Custom Numeric Format Strings. By default, the Parse and TryParse methods can successfully convert strings that contain integral decimal digits only to integer values. They can successfully convert strings that contain integral and fractional decimal digits, group separators, and a decimal separator to floating-point values. The Parse method throws an exception if the operation fails, whereas the TryParse method returns false.

Note

Starting in .NET 7, the numeric types in .NET also implement the System.IParsable<TSelf> interface, which defines the IParsable<TSelf>.Parse and IParsable<TSelf>.TryParse methods.

Parsing and format providers

Typically, the string representations of numeric values differ by culture. Elements of numeric strings, such as currency symbols, group (or thousands) separators, and decimal separators, all vary by culture. Parsing methods either implicitly or explicitly use a format provider that recognizes these culture-specific variations. If no format provider is specified in a call to the Parse or TryParse method, the format provider associated with the current culture (the NumberFormatInfo object returned by the NumberFormatInfo.CurrentInfo property) is used.

A format provider is represented by an IFormatProvider implementation. This interface has a single member, the GetFormat method, whose single parameter is a Type object that represents the type to be formatted. This method returns the object that provides formatting information. .NET supports the following two IFormatProvider implementations for parsing numeric strings:

The following example tries to convert each string in an array to a Double value. It first tries to parse the string by using a format provider that reflects the conventions of the English (United States) culture. If this operation throws a FormatException, it tries to parse the string by using a format provider that reflects the conventions of the French (France) culture.

using System;
using System.Globalization;

public class Example
{
   public static void Main()
   {
      string[] values = { "1,304.16", "$1,456.78", "1,094", "152",
                          "123,45 €", "1 304,16", "Ae9f" };
      double number;
      CultureInfo culture = null;

      foreach (string value in values) {
         try {
            culture = CultureInfo.CreateSpecificCulture("en-US");
            number = Double.Parse(value, culture);
            Console.WriteLine("{0}: {1} --> {2}", culture.Name, value, number);
         }
         catch (FormatException) {
            Console.WriteLine("{0}: Unable to parse '{1}'.",
                              culture.Name, value);
            culture = CultureInfo.CreateSpecificCulture("fr-FR");
            try {
               number = Double.Parse(value, culture);
               Console.WriteLine("{0}: {1} --> {2}", culture.Name, value, number);
            }
            catch (FormatException) {
               Console.WriteLine("{0}: Unable to parse '{1}'.",
                                 culture.Name, value);
            }
         }
         Console.WriteLine();
      }
   }
}
// The example displays the following output:
//    en-US: 1,304.16 --> 1304.16
//
//    en-US: Unable to parse '$1,456.78'.
//    fr-FR: Unable to parse '$1,456.78'.
//
//    en-US: 1,094 --> 1094
//
//    en-US: 152 --> 152
//
//    en-US: Unable to parse '123,45 €'.
//    fr-FR: Unable to parse '123,45 €'.
//
//    en-US: Unable to parse '1 304,16'.
//    fr-FR: 1 304,16 --> 1304.16
//
//    en-US: Unable to parse 'Ae9f'.
//    fr-FR: Unable to parse 'Ae9f'.
Imports System.Globalization

Module Example
    Public Sub Main()
        Dim values() As String = {"1,304.16", "$1,456.78", "1,094", "152",
                                   "123,45 €", "1 304,16", "Ae9f"}
        Dim number As Double
        Dim culture As CultureInfo = Nothing

        For Each value As String In values
            Try
                culture = CultureInfo.CreateSpecificCulture("en-US")
                number = Double.Parse(value, culture)
                Console.WriteLine("{0}: {1} --> {2}", culture.Name, value, number)
            Catch e As FormatException
                Console.WriteLine("{0}: Unable to parse '{1}'.",
                                  culture.Name, value)
                culture = CultureInfo.CreateSpecificCulture("fr-FR")
                Try
                    number = Double.Parse(value, culture)
                    Console.WriteLine("{0}: {1} --> {2}", culture.Name, value, number)
                Catch ex As FormatException
                    Console.WriteLine("{0}: Unable to parse '{1}'.",
                                      culture.Name, value)
                End Try
            End Try
            Console.WriteLine()
        Next
    End Sub
End Module
' The example displays the following output:
'    en-US: 1,304.16 --> 1304.16
'    
'    en-US: Unable to parse '$1,456.78'.
'    fr-FR: Unable to parse '$1,456.78'.
'    
'    en-US: 1,094 --> 1094
'    
'    en-US: 152 --> 152
'    
'    en-US: Unable to parse '123,45 €'.
'    fr-FR: Unable to parse '123,45 €'.
'    
'    en-US: Unable to parse '1 304,16'.
'    fr-FR: 1 304,16 --> 1304.16
'    
'    en-US: Unable to parse 'Ae9f'.
'    fr-FR: Unable to parse 'Ae9f'.

Parsing and NumberStyles Values

The style elements (such as white space, group separators, and decimal separator) that the parse operation can handle are defined by a NumberStyles enumeration value. By default, strings that represent integer values are parsed by using the NumberStyles.Integer value, which permits only numeric digits, leading and trailing white space, and a leading sign. Strings that represent floating-point values are parsed using a combination of the NumberStyles.Float and NumberStyles.AllowThousands values; this composite style permits decimal digits along with leading and trailing white space, a leading sign, a decimal separator, a group separator, and an exponent. By calling an overload of the Parse or TryParse method that includes a parameter of type NumberStyles and setting one or more NumberStyles flags, you can control the style elements that can be present in the string for the parse operation to succeed.

For example, a string that contains a group separator can't be converted to an Int32 value by using the Int32.Parse(String) method. However, the conversion succeeds if you use the NumberStyles.AllowThousands flag, as the following example illustrates.

using System;
using System.Globalization;

public class Example
{
   public static void Main()
   {
      string value = "1,304";
      int number;
      IFormatProvider provider = CultureInfo.CreateSpecificCulture("en-US");
      if (Int32.TryParse(value, out number))
         Console.WriteLine("{0} --> {1}", value, number);
      else
         Console.WriteLine("Unable to convert '{0}'", value);

      if (Int32.TryParse(value, NumberStyles.Integer | NumberStyles.AllowThousands,
                        provider, out number))
         Console.WriteLine("{0} --> {1}", value, number);
      else
         Console.WriteLine("Unable to convert '{0}'", value);
   }
}
// The example displays the following output:
//       Unable to convert '1,304'
//       1,304 --> 1304
Imports System.Globalization

Module Example
    Public Sub Main()
        Dim value As String = "1,304"
        Dim number As Integer
        Dim provider As IFormatProvider = CultureInfo.CreateSpecificCulture("en-US")
        If Int32.TryParse(value, number) Then
            Console.WriteLine("{0} --> {1}", value, number)
        Else
            Console.WriteLine("Unable to convert '{0}'", value)
        End If

        If Int32.TryParse(value, NumberStyles.Integer Or NumberStyles.AllowThousands,
                          provider, number) Then
            Console.WriteLine("{0} --> {1}", value, number)
        Else
            Console.WriteLine("Unable to convert '{0}'", value)
        End If
    End Sub
End Module
' The example displays the following output:
'       Unable to convert '1,304'
'       1,304 --> 1304

Warning

The parse operation always uses the formatting conventions of a particular culture. If you do not specify a culture by passing a CultureInfo or NumberFormatInfo object, the culture associated with the current thread is used.

The following table lists the members of the NumberStyles enumeration and describes the effect that they have on the parsing operation.

NumberStyles value Effect on the string to be parsed
NumberStyles.None Only numeric digits are permitted.
NumberStyles.AllowDecimalPoint The decimal separator and fractional digits are permitted. For integer values, only zero is permitted as a fractional digit. Valid decimal separators are determined by the NumberFormatInfo.NumberDecimalSeparator or NumberFormatInfo.CurrencyDecimalSeparator property.
NumberStyles.AllowExponent The "e" or "E" character can be used to indicate exponential notation. For more information, see NumberStyles.
NumberStyles.AllowLeadingWhite Leading white space is permitted.
NumberStyles.AllowTrailingWhite Trailing white space is permitted.
NumberStyles.AllowLeadingSign A positive or negative sign can precede numeric digits.
NumberStyles.AllowTrailingSign A positive or negative sign can follow numeric digits.
NumberStyles.AllowParentheses Parentheses can be used to indicate negative values.
NumberStyles.AllowThousands The group separator is permitted. The group separator character is determined by the NumberFormatInfo.NumberGroupSeparator or NumberFormatInfo.CurrencyGroupSeparator property.
NumberStyles.AllowCurrencySymbol The currency symbol is permitted. The currency symbol is defined by the NumberFormatInfo.CurrencySymbol property.
NumberStyles.AllowHexSpecifier The string to be parsed is interpreted as a hexadecimal number. It can include the hexadecimal digits 0-9, A-F, and a-f. This flag can be used only to parse integer values.

In addition, the NumberStyles enumeration provides the following composite styles, which include multiple NumberStyles flags.

Composite NumberStyles value Includes members
NumberStyles.Integer Includes the NumberStyles.AllowLeadingWhite, NumberStyles.AllowTrailingWhite, and NumberStyles.AllowLeadingSign styles. This is the default style used to parse integer values.
NumberStyles.Number Includes the NumberStyles.AllowLeadingWhite, NumberStyles.AllowTrailingWhite, NumberStyles.AllowLeadingSign, NumberStyles.AllowTrailingSign, NumberStyles.AllowDecimalPoint, and NumberStyles.AllowThousands styles.
NumberStyles.Float Includes the NumberStyles.AllowLeadingWhite, NumberStyles.AllowTrailingWhite, NumberStyles.AllowLeadingSign, NumberStyles.AllowDecimalPoint, and NumberStyles.AllowExponent styles.
NumberStyles.Currency Includes all styles except NumberStyles.AllowExponent and NumberStyles.AllowHexSpecifier.
NumberStyles.Any Includes all styles except NumberStyles.AllowHexSpecifier.
NumberStyles.HexNumber Includes the NumberStyles.AllowLeadingWhite, NumberStyles.AllowTrailingWhite, and NumberStyles.AllowHexSpecifier styles.

Parsing and Unicode Digits

The Unicode standard defines code points for digits in various writing systems. For example, code points from U+0030 to U+0039 represent the basic Latin digits 0 through 9, code points from U+09E6 to U+09EF represent the Bangla digits 0 through 9, and code points from U+FF10 to U+FF19 represent the Fullwidth digits 0 through 9. However, the only numeric digits recognized by parsing methods are the basic Latin digits 0-9 with code points from U+0030 to U+0039. If a numeric parsing method is passed a string that contains any other digits, the method throws a FormatException.

The following example uses the Int32.Parse method to parse strings that consist of digits in different writing systems. As the output from the example shows, the attempt to parse the basic Latin digits succeeds, but the attempt to parse the Fullwidth, Arabic-Indic, and Bangla digits fails.

using System;

public class Example
{
   public static void Main()
   {
      string value;
      // Define a string of basic Latin digits 1-5.
      value = "\u0031\u0032\u0033\u0034\u0035";
      ParseDigits(value);

      // Define a string of Fullwidth digits 1-5.
      value = "\uFF11\uFF12\uFF13\uFF14\uFF15";
      ParseDigits(value);

      // Define a string of Arabic-Indic digits 1-5.
      value = "\u0661\u0662\u0663\u0664\u0665";
      ParseDigits(value);

      // Define a string of Bangla digits 1-5.
      value = "\u09e7\u09e8\u09e9\u09ea\u09eb";
      ParseDigits(value);
   }

   static void ParseDigits(string value)
   {
      try {
         int number = Int32.Parse(value);
         Console.WriteLine("'{0}' --> {1}", value, number);
      }
      catch (FormatException) {
         Console.WriteLine("Unable to parse '{0}'.", value);
      }
   }
}
// The example displays the following output:
//       '12345' --> 12345
//       Unable to parse '12345'.
//       Unable to parse '١٢٣٤٥'.
//       Unable to parse '১২৩৪৫'.
Module Example
    Public Sub Main()
        Dim value As String
        ' Define a string of basic Latin digits 1-5.
        value = ChrW(&h31) + ChrW(&h32) + ChrW(&h33) + ChrW(&h34) + ChrW(&h35)
        ParseDigits(value)

        ' Define a string of Fullwidth digits 1-5.
        value = ChrW(&hff11) + ChrW(&hff12) + ChrW(&hff13) + ChrW(&hff14) + ChrW(&hff15)
        ParseDigits(value)

        ' Define a string of Arabic-Indic digits 1-5.
        value = ChrW(&h661) + ChrW(&h662) + ChrW(&h663) + ChrW(&h664) + ChrW(&h665)
        ParseDigits(value)

        ' Define a string of Bangla digits 1-5.
        value = ChrW(&h09e7) + ChrW(&h09e8) + ChrW(&h09e9) + ChrW(&h09ea) + ChrW(&h09eb)
        ParseDigits(value)
    End Sub

    Sub ParseDigits(value As String)
        Try
            Dim number As Integer = Int32.Parse(value)
            Console.WriteLine("'{0}' --> {1}", value, number)
        Catch e As FormatException
            Console.WriteLine("Unable to parse '{0}'.", value)
        End Try
    End Sub
End Module
' The example displays the following output:
'       '12345' --> 12345
'       Unable to parse '12345'.
'       Unable to parse '١٢٣٤٥'.
'       Unable to parse '১২৩৪৫'.

See also