ToTitleCase ignore ordinals in C #

Question

I am trying to figure out a way to use toTitleCase to ignore ordinals. It works the way I want for all lines except ordinals (e.g. 1st, 2nd, 3rd becomes 1St, 2Nd, 3Rd).

Any help would be greatly appreciated. A regular expression can be a way to handle this, I'm just not sure how such a regular expression will be created.

Update: here is the solution I used (using John's answer I wrote below the extension method):

public static string ToTitleCaseIgnoreOrdinals(this string text) { string input = System.Globalization.CultureInfo.CurrentCulture.TextInfo.ToTitleCase(text); string result = System.Text.RegularExpressions.Regex.Replace(input, "([0-9]st)|([0-9]th)|([0-9]rd)|([0-9]nd)", new System.Text.RegularExpressions.MatchEvaluator((m) => m.Captures[0].Value.ToLower()), System.Text.RegularExpressions.RegexOptions.IgnoreCase); return result; } 

Answer 1

 string input = System.Globalization.CultureInfo.CurrentCulture.TextInfo.ToTitleCase("hello there, this is the 1st"); string result = System.Text.RegularExpressions.Regex.Replace(input, "([0-9]st)|([0-9]th)|([0-9]rd)|([0-9]nd)", new System.Text.RegularExpressions.MatchEvaluator((m) => { return m.Captures[0].Value.ToLower(); }), System.Text.RegularExpressions.RegexOptions.IgnoreCase); 

Answer 2

You can use regular expressions to check if a string starts with a number before converting to a header, for example:

 if (!Regex.IsMatch(text, @"^\d+")) { CultureInfo.CurrentCulture.TextInfo.toTitleCase(text); } 

Edit: forgot to change the conditional ... change so that it applies to TitleCase if it does NOT match.

2nd edit: added loop to check all words in a sentence:

 string text = "150 east 40th street"; string[] array = text.Split(' '); for (int i = 0; i < array.Length; i++) { if (!Regex.IsMatch(array[i], @"^\d+")) { array[i] = CultureInfo.CurrentCulture.TextInfo.ToTitleCase(array[i]); } } string newText = string.Join(" ",array); 

Answer 3

This will work for these lines, you can override ToTitleCase () using the extension method.

 string s = "1st"; if ( s[0] >= '0' && s[0] <= '9' ) { //this string starts with a number //so don't call ToTitleCase() } else { //call ToTileCase() } 

Answer 4

You can simply use String.Replace (or StringBuilder.Replace ):

 string[] ordinals = { "1St", "2Nd", "3Rd" }; // add all others string text = "This is just sample text which contains some ordinals, the 1st, the 2nd and the third."; var sb = new StringBuilder(CultureInfo.InvariantCulture.TextInfo.ToTitleCase(text)); foreach (string ordinal in ordinals) sb.Replace(ordinal, ordinal.ToLowerInvariant()); text = sb.ToString(); 

it is not at all elegant. This requires you to have an endless list of ordinals in the first line. I guess someone downvoted you.

This is not elegant, but works better than other simple approaches like regex. You want the headwords in a longer text. But only words that are not ordinals. The sequence number is fe 1, 2 or 3 and 31, but not the 31st. Thus, simple regex calculations will not be fast. You also want to use headings like 10m to 10m (where M can be a million reduction).

Therefore, I do not understand why it is so bad to maintain a list of serial numbers.

You can even generate them automatically with an upper limit, for example:

 public static IEnumerable<string> GetTitleCaseOrdinalNumbers() { for (int num = 1; num <= int.MaxValue; num++) { switch (num % 100) { case 11: case 12: case 13: yield return num + "Th"; break; } switch (num % 10) { case 1: yield return num + "St"; break; case 2: yield return num + "Nd"; break; case 3: yield return num + "Rd"; break; default: yield return num + "Th"; break; } } } 

So, if you want to check the first 1000 serial numbers:

 foreach (string ordinal in GetTitleCaseOrdinalNumbers().Take(1000)) sb.Replace(ordinal, ordinal.ToLowerInvariant()); 

---

Update

For what it's worth, here is my attempt to provide an efficient way that really validates words (and not just substrings) and skips ToTitleCase words that really represent ordinal numbers (therefore not 31th , but 31st for example). It also takes care of separator characters that are not white spaces (e.g., periods or commas):

 private static readonly char[] separator = { '.', ',', ';', ':', '-', '(', ')', '\\', '{', '}', '[', ']', '/', '\\', '\'', '"', '"', '?', '!', '|' }; public static bool IsOrdinalNumber(string word) { if (word.Any(char.IsWhiteSpace)) return false; // white-spaces are not allowed if (word.Length < 3) return false; var numericPart = word.TakeWhile(char.IsDigit); string numberText = string.Join("", numericPart); if (numberText.Length == 0) return false; int number; if (!int.TryParse(numberText, out number)) return false; // handle unicode digits which are not really numeric like ۵ string ordinalNumber; switch (number % 100) { case 11: case 12: case 13: ordinalNumber = number + "th"; break; } switch (number % 10) { case 1: ordinalNumber = number + "st"; break; case 2: ordinalNumber = number + "nd"; break; case 3: ordinalNumber = number + "rd"; break; default: ordinalNumber = number + "th"; break; } string checkForOrdinalNum = numberText + word.Substring(numberText.Length); return checkForOrdinalNum.Equals(ordinalNumber, StringComparison.CurrentCultureIgnoreCase); } public static string ToTitleCaseIgnoreOrdinalNumbers(string text, TextInfo info) { if(text.Trim().Length < 3) return info.ToTitleCase(text); int whiteSpaceIndex = FindWhiteSpaceIndex(text, 0, separator); if(whiteSpaceIndex == -1) { if(IsOrdinalNumber(text.Trim())) return text; else return info.ToTitleCase(text); } StringBuilder sb = new StringBuilder(); int wordStartIndex = 0; if(whiteSpaceIndex == 0) { // starts with space, find word wordStartIndex = FindNonWhiteSpaceIndex(text, 1, separator); sb.Append(text.Remove(wordStartIndex)); // append leading spaces } while(wordStartIndex >= 0) { whiteSpaceIndex = FindWhiteSpaceIndex(text, wordStartIndex + 1, separator); string word; if(whiteSpaceIndex == -1) word = text.Substring(wordStartIndex); else word = text.Substring(wordStartIndex, whiteSpaceIndex - wordStartIndex); if(IsOrdinalNumber(word)) sb.Append(word); else sb.Append(info.ToTitleCase(word)); wordStartIndex = FindNonWhiteSpaceIndex(text, whiteSpaceIndex + 1, separator); string whiteSpaces; if(wordStartIndex >= 0) whiteSpaces = text.Substring(whiteSpaceIndex, wordStartIndex - whiteSpaceIndex); else whiteSpaces = text.Substring(whiteSpaceIndex); sb.Append(whiteSpaces); // append spaces between words } return sb.ToString(); } public static int FindWhiteSpaceIndex(string text, int startIndex = 0, params char[] separator) { bool checkSeparator = separator != null && separator.Any(); for (int i = startIndex; i < text.Length; i++) { char c = text[i]; if (char.IsWhiteSpace(c) || (checkSeparator && separator.Contains(c))) return i; } return -1; } public static int FindNonWhiteSpaceIndex(string text, int startIndex = 0, params char[] separator) { bool checkSeparator = separator != null && separator.Any(); for (int i = startIndex; i < text.Length; i++) { char c = text[i]; if (!char.IsWhiteSpace(text[i]) && (!checkSeparator || !separator.Contains(c))) return i; } return -1; } 

Please note that this has not yet been verified, but should give you an idea.

Answer 5

I would break the text up and repeat through the resulting array, skipping things that don't start with a letter.

  using System.Globalization; TextInfo textInfo = new CultureInfo("en-US", false).TextInfo; string[] text = myString.Split(); for(int i = 0; i < text.Length; i++) { //Check for zero-length strings, because these will throw an //index out of range exception in Char.IsLetter if (text[i].Length > 0 && Char.IsLetter(text[i][0])) { text[i] = textInfo.ToTitleCase(text[i]); } }