Over the past couple of weeks I had been writing a utility for a friend of mine that involves parsing a CSV file.  One of the fields in the file was for song names.  Song names, as you know, can contain commas in them, so a simple split on the comma was not viable.

The great thing about CSV files is that when a field itself contains a comma, the entire field is "surrounded, in quotation marks".  This features allows us to write code that splits on the commas while ignoring those between quotes, and adds the results to a List<string>.

Here is the code I wrote to accomplish this.  You will want to read the CSV file line by line and pass each line into the ParseFile(string text) method.

private enum EntryType
        {
            Word,
            Comma,
            Quote
        }

        public IList<string> ParseLine(string text)
        {
            IList<string> vals = new List<string>();
            int i = 0;

            while (i < text.Length)
            {
                vals.Add(GetWord(text, ref i));
            }

            //is last character empty?
            if (text[text.Length - 1].ToString() == ",")
            {
                vals.Add(string.Empty);
            }

            return vals;
        }

        private string GetWord(string text, ref int position)
        {
            EntryType entry = EntryType.Word;
            string word = string.Empty;
            while (position < text.Length)
            {
                string letter = text[position].ToString();
                position++;
                switch (letter)
                {
                    case ",":
                        if (entry == EntryType.Word)
                            return word;
                        else if (entry == EntryType.Quote)
                            word += letter;
                        else if (entry == EntryType.Comma)
                            return word;
                        break;
                    case "\"":
                        if (entry == EntryType.Word || entry == EntryType.Comma)
                            entry = EntryType.Quote;  //open quote
                        else if (entry == EntryType.Quote)
                            return word;  //closing quote
                        break;
                    default:
                        if (entry == EntryType.Word)
                            word += letter;
                        else if (entry == EntryType.Quote)
                            word += letter;
                        else if (entry == EntryType.Comma)
                        {
                            word += letter;
                            entry = EntryType.Word;
                        }
                        break;
                }
            }
            return word;
        }

Of course, a week after I wrote this, my Google Reader presents me with a post called LINQ to CSV using DynamicObject and TextFieldParser.  It looks like I'll try refactoring the utility at some point in the near future.

But before that, I have at least one more post I wish to write about this utility.