fuzzy string compare (check for shorthand matching) C#

Question

I have two lists of string, and I want to extract from each list the index if the string at current index is in the second list(and vice versa), the string cant match exactly or can be a shorthand of another list, for example, consider this two list

List<string> aList = new List<string> { "Id", "PartCode", "PartName", "EquipType" };
List<string> bList = new List<string> { "PartCode", "PartName", "PartShortName", "EquipmentType" };

in the above example, I want from aList the indexes: 1,2,3

and from bList indexes 0,1,3

indexes 1,2 from aList are obvious the string matched completely, but the interesting part are "EquipType" and "EquipmentType" which match becuse "EquipType" is a shorthand of "EquipmentType"

but "PartName" is not a shorthand of "PartShortName" so there indexes are not needed

these is my code

List<string> aList = new List<string> { "Id", "PartCode", "PartName", "EquipType" };// 1, 2 , 3
List<string> bList = new List<string> { "PartCode", "PartName", "PartShortName", "EquipmentType" };//0, 1 ,3 

List<int> alistIndex = new List<int>();
List<int> blistIndex = new List<int>();

for (int i = 0; i < aList.Count; i++)           
{
    string a = aList[i];
    for (int j = 0; j < bList.Count(); j++)               
    {
        string b = bList[j];

        string bigger, smaller;
        int biggerCount, smallerCount;
        if (a.Length > b.Length)
        {
            bigger = a; smaller = b;
            biggerCount = a.Length ; smallerCount = b.Length ;    
        }
        else
        {
            bigger = b; smaller = a;
            biggerCount = b.Length; smallerCount = a.Length ;
        }

        int countCheck = 0;
        for (int k = 0; k < biggerCount; k++)
        {
            if (smaller.Length != countCheck)
            {
                if (bigger[k] == smaller[countCheck])
                    countCheck++;
             }
         }

        if (countCheck == smaller.Length)
        {
            alistIndex.Add(i);
            blistIndex.Add(j);
            res = true;
            break;
        }
        else
            res = false;  
    }
}

alistIndex.ForEach(i => Console.Write(i));
Console.WriteLine(Environment.NewLine);
blistIndex.ForEach(i => Console.Write(i));
Console.ReadKey();

the above code works just fine and looks very similar to this solution

but if change the order of the second list like so

 List<string> bList = new List<string> { "PartCode", "PartShortName", "PartName", "EquipmentType" };

i will get index 0, 1 and 3 (but i want 0 2 and 3)

should i check the distance for every pair and return the lowest? or should i work i a different method

Thanks

p.s i also found this GitHub, but i don't know if it will do the trick for me

Answer 1

I do feel that what you are trying to do is a bad idea... Id is the abbreviation of Idiotic, just to give an example :-) Still... I wanted to do some experiments on Unicode.

Now, this code will split words on uppercase letters. PartName is Part + Name because the N is uppercase. It doesn't support ID as Identifier (because it should be IDentifier) but it does support NSA as NotSuchAgency :-) So full acronyms are ok, while FDA isn't equivalent to FoodAndDrugAdministration, so acronyms with conjunctions are KO.

public static bool ShorthandCompare(string str1, string str2)
{
    if (str1 == null)
    {
        throw new ArgumentNullException(nameof(str1));
    }

    if (str2 == null)
    {
        throw new ArgumentNullException(nameof(str2));
    }

    if (str1 == string.Empty)
    {
        return str2 == string.Empty;
    }

    if (object.ReferenceEquals(str1, str2))
    {
        return true;
    }

    var ee1 = StringInfo.GetTextElementEnumerator(str1);
    var ee2 = StringInfo.GetTextElementEnumerator(str2);

    bool eos1, eos2 = true;

    while ((eos1 = ee1.MoveNext()) && (eos2 = ee2.MoveNext()))
    {
        string ch1 = ee1.GetTextElement(), ch2 = ee2.GetTextElement();

        // The string.Compare does some nifty tricks with unicode
        // like string.Compare("ì", "i\u0300") == 0
        if (string.Compare(ch1, ch2) == 0)
        {
            continue;
        }

        UnicodeCategory uc1 = char.GetUnicodeCategory(ch1, 0);
        UnicodeCategory uc2 = char.GetUnicodeCategory(ch2, 0);

        if (uc1 == UnicodeCategory.UppercaseLetter)
        {
            while (uc2 != UnicodeCategory.UppercaseLetter && (eos2 = ee2.MoveNext()))
            {
                ch2 = ee2.GetTextElement();
                uc2 = char.GetUnicodeCategory(ch2, 0);
            }

            if (!eos2 || string.Compare(ch1, ch2) != 0)
            {
                return false;
            }

            continue;
        }
        else if (uc2 == UnicodeCategory.UppercaseLetter)
        {
            while (uc1 != UnicodeCategory.UppercaseLetter && (eos1 = ee1.MoveNext()))
            {
                ch1 = ee1.GetTextElement();
                uc1 = char.GetUnicodeCategory(ch1, 0);
            }

            if (!eos1 || string.Compare(ch1, ch2) != 0)
            {
                return false;
            }

            continue;
        }

        // We already know they are different!
        return false;
    }

    if (eos1)
    {
        while (ee1.MoveNext())
        {
            string ch1 = ee1.GetTextElement();
            UnicodeCategory uc1 = char.GetUnicodeCategory(ch1, 0);

            if (uc1 == UnicodeCategory.UppercaseLetter)
            {
                return false;
            }
        }
    }
    else if (eos2)
    {
        while (ee2.MoveNext())
        {
            string ch2 = ee2.GetTextElement();
            UnicodeCategory uc2 = char.GetUnicodeCategory(ch2, 0);

            if (uc2 == UnicodeCategory.UppercaseLetter)
            {
                return false;
            }
        }
    }

    return true;
}

and then

List<string> aList = new List<string> { "Id", "PartCode", "PartName", "EquipType" };
List<string> bList = new List<string> { "PartCode", "PartName", "PartShortName", "EquipmentType" };

List<List<int>> matches = new List<List<int>>();

for (int i = 0; i < aList.Count; i++)
{
    var lst = new List<int>();
    matches.Add(lst);

    for (int j = 0; j < bList.Count; j++)
    {
        if (ShorthandCompare(aList[i], bList[j]))
        {
            lst.Add(j);
        }
    }
}

Note that the result is a List<List<int>>, because you could have multiple matches for a single word of aList!

Now... the interesting part of the ShorthandCompare is that it tries to be "intelligent" and handle non-BMP Unicode characters (through the use of StringInfo.GetTextElementEnumerator) and handle decomposed Unicode characters (the ì character can be obtained in Unicode through i + \u0300, that is its dieresis). It does it through the use of string.Compare that, differently than string.Equals, is Unicode-aware (string.CompareOrdinal is more similar to string.Equals and not Unicode-aware).

bool cmp1 = ShorthandCompare("IdìoLe\u0300ss", "Idi\u0300oticLèsser"); // true