Add support for Unicode digits

tompazourek · Apr 14, 2024 · c303896 · c303896
1 parent 6ec645d
commit c303896
Show file tree

Hide file tree

Showing 2 changed files with 78 additions and 18 deletions.
diff --git a/src/NaturalSort.Extension/NaturalSortComparer.cs b/src/NaturalSort.Extension/NaturalSortComparer.cs
@@ -125,9 +125,26 @@ public int Compare(string str1, string str2)
                     var digit1 = i < paddingLength1 ? paddingChar : str1[startIndex1 + i - paddingLength1];
                     var digit2 = i < paddingLength2 ? paddingChar : str2[startIndex2 + i - paddingLength2];
 
-                    var digitCompare = digit1.CompareTo(digit2);
-                    if (digitCompare != 0)
-                        return digitCompare;
+                    if (digit1 is >= '0' and <= '9' && digit2 is >= '0' and <= '9')
+                    {
+                        // both digits are ordinary 0 to 9
+                        var digitCompare = digit1.CompareTo(digit2);
+                        if (digitCompare != 0)
+                            return digitCompare;
+                    }
+                    else
+                    {
+                        // one or both digits is unicode, compare parsed numeric values, and only if they are same, compare as char
+                        var digitNumeric1 = char.GetNumericValue(digit1);
+                        var digitNumeric2 = char.GetNumericValue(digit2);
+                        var digitNumericCompare = digitNumeric1.CompareTo(digitNumeric2);
+                        if (digitNumericCompare != 0)
+                            return digitNumericCompare;
+
+                        var digitCompare = digit1.CompareTo(digit2);
+                        if (digitCompare != 0)
+                            return digitCompare;
+                    }
                 }
 
                 // if the numbers are equal, we compare how much we padded the strings
@@ -164,19 +181,54 @@ public int Compare(string str1, string str2)
     }
 
     private static byte GetTokenFromChar(char c)
-        => c >= 'a'
-            ? c <= 'z'
-                ? TokenLetters
-                : c < 128
-                    ? TokenOther
-                    : char.IsLetter(c)
-                        ? TokenLetters
-                        : TokenOther
-            : c >= 'A'
-                ? c <= 'Z'
-                    ? TokenLetters
-                    : TokenOther
-                : c is >= '0' and <= '9'
-                    ? TokenDigits
-                    : TokenOther;
+    {
+        if (c >= 'a')
+        {
+            if (c <= 'z')
+            {
+                return TokenLetters;
+            }
+            else if (c < 128)
+            {
+                return TokenOther;
+            }
+            else if (char.IsLetter(c))
+            {
+                return TokenLetters;
+            }
+            else if (char.IsDigit(c))
+            {
+                return TokenDigits;
+            }
+            else
+            {
+                return TokenOther;
+            }
+        }
+        else
+        {
+            if (c >= 'A')
+            {
+                if (c <= 'Z')
+                {
+                    return TokenLetters;
+                }
+                else
+                {
+                    return TokenOther;
+                }
+            }
+            else
+            {
+                if (c is >= '0' and <= '9')
+                {
+                    return TokenDigits;
+                }
+                else
+                {
+                    return TokenOther;
+                }
+            }
+        }
+    }
 }
diff --git a/tests/NaturalSort.Extension.Tests/NaturalSortComparerTests.cs b/tests/NaturalSort.Extension.Tests/NaturalSortComparerTests.cs
@@ -57,6 +57,14 @@ private static void RunTests_StringComparisonOrdinalIgnoreCase(string[] input, s
         new[] { "x", "x x", "x!x", "x#x", "x%x", "x&x", "x(x", "x)x", "x,x", "x.x", "x;x", "x@x", "x[x", "x]x", "x^x", "x_x", "x{x", "x}x", "x~x", "x0x", "x1x", "x2x", "x3x", "x4x", "x5x", "x6x", "x7x", "x8x", "x9x", "xAx", "xBx", "xCx", "xDx", "xEx", "xFx", "xGx", "xHx", "xIx", "xJx", "xKx", "xLx", "xMx", "xNx", "xOx", "xPx", "xQx", "xRx", "xSx", "xTx", "xUx", "xVx", "xWx", "xx", "xXx", "xYx", "xZx" },
         new[] { "x", "x x", "x!x", "x#x", "x%x", "x&x", "x(x", "x)x", "x,x", "x.x", "x;x", "x@x", "x[x", "x]x", "x^x", "x_x", "x{x", "x}x", "x~x", "x0x", "x1x", "x2x", "x3x", "x4x", "x5x", "x6x", "x7x", "x8x", "x9x", "xAx", "xBx", "xCx", "xDx", "xEx", "xFx", "xGx", "xHx", "xIx", "xJx", "xKx", "xLx", "xMx", "xNx", "xOx", "xPx", "xQx", "xRx", "xSx", "xTx", "xUx", "xVx", "xWx", "xx", "xXx", "xYx", "xZx" }
     )]
+    [InlineData(
+        new[] { "A33", "A3", "A11", "A10", "Z", "A2", "A", "A੨", "A22", "A੨੨" },
+        new[] { "A", "A2", "A੨", "A3", "A10", "A11", "A22", "A੨੨", "A33", "Z" }
+    )]
+    [InlineData(
+        new[] { "A", "A10", "A11", "Z", "A੨", "A੨੨" },
+        new[] { "A", "A੨", "A10", "A11", "A੨੨", "Z" }
+    )]
     public void WindowsExplorer(string[] input, string[] expected) => RunTests(input, expected);
 
     /// <remarks>