如何在C＃中替换字符串中的多个子字符串？

Question

I have to replace multiple substrings from a string (max length 32 of input string). I have a big dictionary which can have millions of items as a key-value pair. I need to check for each word if this word is present in the dictionary and replace with the respective value if present in the dictionary. The input string can have multiple trailing spaces.

This method is being called millions of time, due to this, it's affecting the performance badly.

Is there any scope of optimization in the code or some other better way to do this.

public static string RandomValueCompositeField(object objInput, Dictionary<string, string> g_rawValueRandomValueMapping) {

if (objInput == null)
    return null;

string input = objInput.ToString();
if (input == "")
    return input;

//List<string> ls = new List<string>();
int count = WhiteSpaceAtEnd(input);
foreach (string data in input.Substring(0, input.Length - count).Split(' ')) {

    try {
        string value;
        gs_dictRawValueRandomValueMapping.TryGetValue(data, out value);
        if (value != null) {

            //ls.Add(value.TrimEnd());
            input = input.Replace(data, value);
        }
        else {
            //ls.Add(data);
        }
    }
    catch(Exception ex) {

    }

}

//if (count > 0)
//    input = input + new string(' ', count);
    //ls.Add(new string(' ', count));

return input;
}

EDIT:

I missed one important thing in the question. substring can occur only once inthe input string. Dictionay key and value have same number of characters.

Answer 1

Here's a method that will take an input string and will build a new string by finding "words" (any consecutive non-whitespace) and then checking if that word is in a dictionary and replacing it with the corresponding value if found. This will fix the issues of Replace doing replacements on "sub-words" (if you have "hello hell" and you want to replace "hell" with "heaven" and you don't want it to give you "heaveno heaven"). It also fixes the issue of swapping. For example if you want to replace "yes" with "no" and "no" with "yes" in "yes no" you don't want it to first turn that into "no no" and then into "yes yes".

public string ReplaceWords(string input, Dictionary<string, string> replacements)
{
    var builder = new StringBuilder();
    int wordStart = -1;
    int wordLength = 0;
    for(int i = 0; i < input.Length; i++)
    {
        // If the current character is white space check if we have a word to replace
        if(char.IsWhiteSpace(input[i]))
        {
            // If wordStart is not -1 then we have hit the end of a word
            if(wordStart >= 0)
            {
                // get the word and look it up in the dictionary
                // if found use the replacement, if not keep the word.
                var word = input.Substring(wordStart, wordLength);
                if(replacements.TryGetValue(word, out var replace))
                {
                    builder.Append(replace);
                }
                else
                {
                    builder.Append(word);
                }
            }

            // Make sure to reset the start and length
            wordStart = -1;
            wordLength = 0;

            // append whatever whitespace was found.
            builder.Append(input[i]);
        }
        // If this isn't whitespace we set wordStart if it isn't already set
        // and just increment the length.
        else
        {
            if(wordStart == -1) wordStart = i;
            wordLength++;
        }
    }

    // If wordStart is not -1 then we have a trailing word we need to check.
    if(wordStart >= 0)
    {
        var word = input.Substring(wordStart, wordLength);
        if(replacements.TryGetValue(word, out var replace))
        {
            builder.Append(replace);
        }
        else
        {
            builder.Append(word);
        }
    }

    return builder.ToString();
}