142

Given

2,1016,7/31/2008 14:22,Geoff Dalgas,6/5/2011 22:21,http://stackoverflow.com,"Corvallis, OR",7679,351,81,b437f461b3fd27387c5d8ab47a293d35,34

How to use C# to split the above information into strings as follows:

2
1016
7/31/2008 14:22
Geoff Dalgas
6/5/2011 22:21
http://stackoverflow.com
Corvallis, OR
7679
351
81
b437f461b3fd27387c5d8ab47a293d35
34

As you can see one of the column contains , <= (Corvallis, OR)

Based on C# Regex Split - commas outside quotes

string[] result = Regex.Split(samplestring, ",(?=(?:[^\"]*\"[^\"]*\")*[^\"]*$)");
Dour High Arch
  • 21,513
  • 29
  • 75
  • 90
q0987
  • 34,938
  • 69
  • 242
  • 387
  • 1
    Though in Java, similar Question : http://stackoverflow.com/questions/1757065/java-splitting-a-comma-separated-string-but-ignoring-commas-in-quotes – Saurabh Gokhale Jul 01 '11 at 02:28
  • 1
    Using a regex to do this is bad advice. The .NET Framework already has built-in support to parse CSV. See this answer which is the one you should accept. Otherwise I'll close this as a dupe of http://stackoverflow.com/questions/3147836/c-regex-split-commas-outside-quotes which is just as equally wrong. – Kev Jul 01 '11 at 18:13
  • 1
    Can you please elaborate what is .NET's in-built support for parsing CSV files with embedded commas? Are you referring to the Microsoft.VisualBasic.FileIO.TextFieldParser class? – AllSolutions Sep 24 '16 at 21:50
  • Does this answer your question? [Reading CSV files using C#](https://stackoverflow.com/questions/3507498/reading-csv-files-using-c-sharp) – Dour High Arch Nov 02 '20 at 15:48
  • Just another `Regex` based solution [here](https://stackoverflow.com/a/43284021/5770014) with some nice example. – minus one Dec 07 '21 at 19:21

9 Answers9

213

Use the Microsoft.VisualBasic.FileIO.TextFieldParser class. This will handle parsing a delimited file, TextReader or Stream where some fields are enclosed in quotes and some are not.

For example:

using Microsoft.VisualBasic.FileIO;

string csv = "2,1016,7/31/2008 14:22,Geoff Dalgas,6/5/2011 22:21,http://stackoverflow.com,\"Corvallis, OR\",7679,351,81,b437f461b3fd27387c5d8ab47a293d35,34";

TextFieldParser parser = new TextFieldParser(new StringReader(csv));

// You can also read from a file
// TextFieldParser parser = new TextFieldParser("mycsvfile.csv");

parser.HasFieldsEnclosedInQuotes = true;
parser.SetDelimiters(",");

string[] fields;

while (!parser.EndOfData)
{
    fields = parser.ReadFields();
    foreach (string field in fields)
    {
        Console.WriteLine(field);
    }
} 

parser.Close();

This should result in the following output:

2
1016
7/31/2008 14:22
Geoff Dalgas
6/5/2011 22:21
http://stackoverflow.com
Corvallis, OR
7679
351
81
b437f461b3fd27387c5d8ab47a293d35
34

See Microsoft.VisualBasic.FileIO.TextFieldParser for more information.

You need to add a reference to Microsoft.VisualBasic in the Add References .NET tab.

M.Babcock
  • 18,753
  • 6
  • 54
  • 84
Tim
  • 28,212
  • 8
  • 63
  • 76
  • 13
    Dude, thank you very much for this solution, I have about 500K+ rows of CSV data that I need to load into a table and it loaded with commas contained inside of quotes. I owe you an adult beverage of your choice if our paths ever cross. – Mark Kram Jul 30 '14 at 17:56
  • @tim i used this, and notices its skipping all even line numbers , only processing the odd line numbers in a file that has 1050 lines. any ideas? – Smith Jun 03 '15 at 15:40
  • @Smith - Without seeing your code or sample input I have no idea. I suggest posting a new question. Maybe the file is missing a carriage return or other end-of-line marker on the even lines? – Tim Jun 03 '15 at 16:02
  • I didn't even know about this library until I saw this - thanks! If anyone else wants an example that parses a whole CSV file, see this SO answer: http://stackoverflow.com/a/3508572/3105807 – Amy Barrett Jul 01 '16 at 10:02
  • 3
    Can we lynch Microsoft for not providing a constructor that takes a string so we have to jump through the hoop of converting it to a stream first?? Otherwise, nice answer. – Loren Pechtel Dec 10 '16 at 01:16
  • I have the `csv` as a string variable which I am exporting as `.csv`. So when do I process the function you mentioned? Before saving it? – Si8 Jan 27 '17 at 20:29
  • @LorenPechtel it does support stringpath as a constructor... he just uses stream because he loads the string from memory (code). – Frode F. Mar 29 '18 at 19:03
  • This VB library is amazingly robust and solved more CSV parsing problems for me than just what was asked in the question. (Why do we have to reach into VB to get this??) – HerrimanCoder Aug 31 '18 at 16:13
  • Hi, this works for this case but not for this one: (any ideas?) string test = "\"Car\", \"Cat\", \"1234\", \",\", \"123\", \"House\"s Color\""; – Stavros Koureas Oct 12 '18 at 13:53
  • Only someone who has run every conceivable test case against every single input form can recommend a library. If you haven't done this, you're recommending something untested. Or maybe you think MS is magically infallible. To me, MS produces junk api's and framework code .. –  Nov 02 '18 at 22:15
  • @sln - Then don't use it. I've used it a number of times and it suited my needs. OP asked a question, I gave an answer that suited him. Do you personally thoroughly test for ever conceivable case for the library's and frameworks you use and recommend? If you don't like MS's stuff, then don't use it - plenty of alternatives out there :) – Tim Nov 03 '18 at 05:18
  • Hey, no problem. I use regex to parse csv and, why yes I do personally test for every conceivable case. Fwiw, I have many years using MFC in production code. Maybe you think MS has magically changed the way it puts together code, but I assure you it hasn't. –  Nov 03 '18 at 11:34
  • you made my day after 3 days try... this works – Shiwanka Chathuranga May 07 '21 at 16:15
  • Can't believe I've worked with CSV files for so long and not known about this class! Truly a massive effort-saver, thank you! – LordPupazz May 16 '22 at 10:06
  • I think the regex parser is a cleaner way around this problem. Nice option though. – Joel Sep 15 '22 at 18:49
92

It is so much late but this can be helpful for someone. We can use RegEx as bellow.

Regex CSVParser = new Regex(",(?=(?:[^\"]*\"[^\"]*\")*(?![^\"]*\"))");
String[] Fields = CSVParser.Split(Test);
Husen
  • 1,541
  • 10
  • 14
5

I see that if you paste csv delimited text in Excel and do a "Text to Columns", it asks you for a "text qualifier". It's defaulted to a double quote so that it treats text within double quotes as literal. I imagine that Excel implements this by going one character at a time, if it encounters a "text qualifier", it keeps going to the next "qualifier". You can probably implement this yourself with a for loop and a boolean to denote if you're inside literal text.

public string[] CsvParser(string csvText)
{
    List<string> tokens = new List<string>();

    int last = -1;
    int current = 0;
    bool inText = false;

    while(current < csvText.Length)
    {
        switch(csvText[current])
        {
            case '"':
                inText = !inText; break;
            case ',':
                if (!inText) 
                {
                    tokens.Add(csvText.Substring(last + 1, (current - last)).Trim(' ', ',')); 
                    last = current;
                }
                break;
            default:
                break;
        }
        current++;
    }

    if (last != csvText.Length - 1) 
    {
        tokens.Add(csvText.Substring(last+1).Trim());
    }

    return tokens.ToArray();
}
Roly
  • 1,516
  • 1
  • 15
  • 26
4

You could split on all commas that do have an even number of quotes following them.

You would also like to view at the specf for CSV format about handling comma's.

Useful Link : C# Regex Split - commas outside quotes

Community
  • 1
  • 1
Saurabh Gokhale
  • 53,625
  • 36
  • 139
  • 164
  • 3
    @q0987 - this is not the right answer. There is built in support for this in the Framework: http://stackoverflow.com/questions/6542996/how-to-split-csv-whose-columns-may-contain/6543418#6543418 – Kev Jul 01 '11 at 18:10
  • @Kev - That is not correct, either. The answer you linked to is for the `Microsoft.VisualBasic.FileIO.TextFieldParser` class, which is NOT built into the .NET Framework. This is why you have to reference the Microsoft.VisualBasic assembly in order to use it your project(s). – Blair Allen Nov 11 '20 at 21:01
  • @BlairAllen Just because it lives in a different namespace and assembly doesn't mean it's not a shipping-from-Microsoft .NET Framework feature. Now, if I'd said it was part of the Base Class Library then you might have a point. – Kev Nov 12 '20 at 22:14
3

It is a tricky matter to parse .csv files when the .csv file could be either comma separated strings, comma separated quoted strings, or a chaotic combination of the two. The solution I came up with allows for any of the three possibilities.

I created a method, ParseCsvRow() which returns an array from a csv string. I first deal with double quotes in the string by splitting the string on double quotes into an array called quotesArray. Quoted string .csv files are only valid if there is an even number of double quotes. Double quotes in a column value should be replaced with a pair of double quotes (This is Excel's approach). As long as the .csv file meets these requirements, you can expect the delimiter commas to appear only outside of pairs of double quotes. Commas inside of pairs of double quotes are part of the column value and should be ignored when splitting the .csv into an array.

My method will test for commas outside of double quote pairs by looking only at even indexes of the quotesArray. It also removes double quotes from the start and end of column values.

    public static string[] ParseCsvRow(string csvrow)
    {
        const string obscureCharacter = "ᖳ";
        if (csvrow.Contains(obscureCharacter)) throw new Exception("Error: csv row may not contain the " + obscureCharacter + " character");

        var unicodeSeparatedString = "";

        var quotesArray = csvrow.Split('"');  // Split string on double quote character
        if (quotesArray.Length > 1)
        {
            for (var i = 0; i < quotesArray.Length; i++)
            {
                // CSV must use double quotes to represent a quote inside a quoted cell
                // Quotes must be paired up
                // Test if a comma lays outside a pair of quotes.  If so, replace the comma with an obscure unicode character
                if (Math.Round(Math.Round((decimal) i/2)*2) == i)
                {
                    var s = quotesArray[i].Trim();
                    switch (s)
                    {
                        case ",":
                            quotesArray[i] = obscureCharacter;  // Change quoted comma seperated string to quoted "obscure character" seperated string
                            break;
                        default:
                            if (s.All(chars => chars == ','))
                            {
                                quotesArray[i] = "";
                                for (int j = 0; j < s.Count(); j++)
                                {
                                    quotesArray[i] += obscureCharacter;
                                } 
                            }
                            break;
                    }
                }
                // Build string and Replace quotes where quotes were expected.
                unicodeSeparatedString += (i > 0 ? "\"" : "") + quotesArray[i].Trim();
            }
        }
        else
        {
            // String does not have any pairs of double quotes.  It should be safe to just replace the commas with the obscure character
            unicodeSeparatedString = csvrow.Replace(",", obscureCharacter);
        }

        var csvRowArray = unicodeSeparatedString.Split(obscureCharacter[0]); 

        for (var i = 0; i < csvRowArray.Length; i++)
        {
            var s = csvRowArray[i].Trim();
            if (s.StartsWith("\"") && s.EndsWith("\""))
            {
                csvRowArray[i] = s.Length > 2 ? s.Substring(1, s.Length - 2) : "";  // Remove start and end quotes.
            }
        }
        
        return csvRowArray;
    }

One downside of my approach is the way I temporarily replace delimiter commas with an obscure unicode character. This character needs to be so obscure, it would never show up in your .csv file. You may want to put more handling around this.

Vi. Bu.
  • 23
  • 4
Jason Williams
  • 2,740
  • 28
  • 36
3

Use a library like LumenWorks to do your CSV reading. It'll handle fields with quotes in them and will likely overall be more robust than your custom solution by virtue of having been around for a long time.

Adam Lear
  • 38,111
  • 12
  • 81
  • 101
3

This question and its duplicates have a lot of answers. I tried this one that looked promising, but found some bugs in it. I heavily modified it so that it would pass all of my tests.

    /// <summary>
    /// Returns a collection of strings that are derived by splitting the given source string at
    /// characters given by the 'delimiter' parameter.  However, a substring may be enclosed between
    /// pairs of the 'qualifier' character so that instances of the delimiter can be taken as literal
    /// parts of the substring.  The method was originally developed to split comma-separated text
    /// where quotes could be used to qualify text that contains commas that are to be taken as literal
    /// parts of the substring.  For example, the following source:
    ///     A, B, "C, D", E, "F, G"
    /// would be split into 5 substrings:
    ///     A
    ///     B
    ///     C, D
    ///     E
    ///     F, G
    /// When enclosed inside of qualifiers, the literal for the qualifier character may be represented
    /// by two consecutive qualifiers.  The two consecutive qualifiers are distinguished from a closing
    /// qualifier character.  For example, the following source:
    ///     A, "B, ""C"""
    /// would be split into 2 substrings:
    ///     A
    ///     B, "C"
    /// </summary>
    /// <remarks>Originally based on: https://stackoverflow.com/a/43284485/2998072</remarks>
    /// <param name="source">The string that is to be split</param>
    /// <param name="delimiter">The character that separates the substrings</param>
    /// <param name="qualifier">The character that is used (in pairs) to enclose a substring</param>
    /// <param name="toTrim">If true, then whitespace is removed from the beginning and end of each
    /// substring.  If false, then whitespace is preserved at the beginning and end of each substring.
    /// </param>
    public static List<String> SplitQualified(this String source, Char delimiter, Char qualifier,
                                Boolean toTrim)
    {
        // Avoid throwing exception if the source is null
        if (String.IsNullOrEmpty(source))
            return new List<String> { "" };

        var results = new List<String>();
        var result = new StringBuilder();
        Boolean inQualifier = false;

        // The algorithm is designed to expect a delimiter at the end of each substring, but the
        // expectation of the caller is that the final substring is not terminated by delimiter.
        // Therefore, we add an artificial delimiter at the end before looping through the source string.
        String sourceX = source + delimiter;

        // Loop through each character of the source
        for (var idx = 0; idx < sourceX.Length; idx++)
        {
            // If current character is a delimiter
            // (except if we're inside of qualifiers, we ignore the delimiter)
            if (sourceX[idx] == delimiter && inQualifier == false)
            {
                // Terminate the current substring by adding it to the collection
                // (trim if specified by the method parameter)
                results.Add(toTrim ? result.ToString().Trim() : result.ToString());
                result.Clear();
            }
            // If current character is a qualifier
            else if (sourceX[idx] == qualifier)
            {
                // ...and we're already inside of qualifier
                if (inQualifier)
                {
                    // check for double-qualifiers, which is escape code for a single
                    // literal qualifier character.
                    if (idx + 1 < sourceX.Length && sourceX[idx + 1] == qualifier)
                    {
                        idx++;
                        result.Append(sourceX[idx]);
                        continue;
                    }
                    // Since we found only a single qualifier, that means that we've
                    // found the end of the enclosing qualifiers.
                    inQualifier = false;
                    continue;
                }
                else
                    // ...we found an opening qualifier
                    inQualifier = true;
            }
            // If current character is neither qualifier nor delimiter
            else
                result.Append(sourceX[idx]);
        }

        return results;
    }

Here are the test methods to prove that it works:

    [TestMethod()]
    public void SplitQualified_00()
    {
        // Example with no substrings
        String s = "";
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_00A()
    {
        // just a single delimiter
        String s = ",";
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "", "" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_01()
    {
        // Example with no whitespace or qualifiers
        String s = "1,2,3,1,2,3";
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_02()
    {
        // Example with whitespace and no qualifiers
        String s = " 1, 2 ,3,  1  ,2\t,   3   ";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_03()
    {
        // Example with whitespace and no qualifiers
        String s = " 1, 2 ,3,  1  ,2\t,   3   ";
        // whitespace should be preserved
        var substrings = s.SplitQualified(',', '"', false);
        CollectionAssert.AreEquivalent(
            new List<String> { " 1", " 2 ", "3", "  1  ", "2\t", "   3   " },
            substrings);
    }
    [TestMethod()]
    public void SplitQualified_04()
    {
        // Example with no whitespace and trivial qualifiers.
        String s = "1,\"2\",3,1,2,\"3\"";
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);

        s = "\"1\",\"2\",3,1,\"2\",3";
        substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_05()
    {
        // Example with no whitespace and qualifiers that enclose delimiters
        String s = "1,\"2,2a\",3,1,2,\"3,3a\"";
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2,2a", "3", "1", "2", "3,3a" },
                                substrings);

        s = "\"1,1a\",\"2,2b\",3,1,\"2,2c\",3";
        substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1,1a", "2,2b", "3", "1", "2,2c", "3" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_06()
    {
        // Example with qualifiers enclosing whitespace but no delimiter
        String s = "\" 1 \",\"2 \",3,1,2,\"\t3\t\"";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_07()
    {
        // Example with qualifiers enclosing whitespace but no delimiter
        String s = "\" 1 \",\"2 \",3,1,2,\"\t3\t\"";
        // whitespace should be preserved
        var substrings = s.SplitQualified(',', '"', false);
        CollectionAssert.AreEquivalent(new List<String> { " 1 ", "2 ", "3", "1", "2", "\t3\t" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_08()
    {
        // Example with qualifiers enclosing whitespace but no delimiter; also whitespace btwn delimiters
        String s = "\" 1 \", \"2 \"  ,  3,1, 2 ,\"  3  \"";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_09()
    {
        // Example with qualifiers enclosing whitespace but no delimiter; also whitespace btwn delimiters
        String s = "\" 1 \", \"2 \"  ,  3,1, 2 ,\"  3  \"";
        // whitespace should be preserved
        var substrings = s.SplitQualified(',', '"', false);
        CollectionAssert.AreEquivalent(new List<String> { " 1 ", " 2   ", "  3", "1", " 2 ", "  3  " },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_10()
    {
        // Example with qualifiers enclosing whitespace and delimiter
        String s = "\" 1 \",\"2 , 2b \",3,1,2,\"  3,3c  \"";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2 , 2b", "3", "1", "2", "3,3c" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_11()
    {
        // Example with qualifiers enclosing whitespace and delimiter; also whitespace btwn delimiters
        String s = "\" 1 \", \"2 , 2b \"  ,  3,1, 2 ,\"  3,3c  \"";
        // whitespace should be preserved
        var substrings = s.SplitQualified(',', '"', false);
        CollectionAssert.AreEquivalent(new List<String> { " 1 ", " 2 , 2b   ", "  3", "1", " 2 ", "  3,3c  " },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_12()
    {
        // Example with tab characters between delimiters
        String s = "\t1,\t2\t,3,1,\t2\t,\t3\t";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_13()
    {
        // Example with newline characters between delimiters
        String s = "\n1,\n2\n,3,1,\n2\n,\n3\n";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_14()
    {
        // Example with qualifiers enclosing whitespace and delimiter, plus escaped qualifier
        String s = "\" 1 \",\"\"\"2 , 2b \"\"\",3,1,2,\"  \"\"3,3c  \"";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "\"2 , 2b \"", "3", "1", "2", "\"3,3c" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_14A()
    {
        // Example with qualifiers enclosing whitespace and delimiter, plus escaped qualifier
        String s = "\"\"\"1\"\"\"";
        // whitespace should be removed
        var substrings = s.SplitQualified(',', '"', true);
        CollectionAssert.AreEquivalent(new List<String> { "\"1\"" },
                                substrings);
    }


    [TestMethod()]
    public void SplitQualified_15()
    {
        // Instead of comma-delimited and quote-qualified, use pipe and hash

        // Example with no whitespace or qualifiers
        String s = "1|2|3|1|2,2f|3";
        var substrings = s.SplitQualified('|', '#', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2", "3", "1", "2,2f", "3" }, substrings);
    }
    [TestMethod()]
    public void SplitQualified_16()
    {
        // Instead of comma-delimited and quote-qualified, use pipe and hash

        // Example with qualifiers enclosing whitespace and delimiter
        String s = "# 1 #|#2 | 2b #|3|1|2|#  3|3c  #";
        // whitespace should be removed
        var substrings = s.SplitQualified('|', '#', true);
        CollectionAssert.AreEquivalent(new List<String> { "1", "2 | 2b", "3", "1", "2", "3|3c" },
                                substrings);
    }
    [TestMethod()]
    public void SplitQualified_17()
    {
        // Instead of comma-delimited and quote-qualified, use pipe and hash

        // Example with qualifiers enclosing whitespace and delimiter; also whitespace btwn delimiters
        String s = "# 1 #| #2 | 2b #  |  3|1| 2 |#  3|3c  #";
        // whitespace should be preserved
        var substrings = s.SplitQualified('|', '#', false);
        CollectionAssert.AreEquivalent(new List<String> { " 1 ", " 2 | 2b   ", "  3", "1", " 2 ", "  3|3c  " },
                                substrings);
    }
Tony Pulokas
  • 465
  • 5
  • 12
1

I had a problem with a CSV that contains fields with a quote character in them, so using the TextFieldParser, I came up with the following:

private static string[] parseCSVLine(string csvLine)
{
  using (TextFieldParser TFP = new TextFieldParser(new MemoryStream(Encoding.UTF8.GetBytes(csvLine))))
  {
    TFP.HasFieldsEnclosedInQuotes = true;
    TFP.SetDelimiters(",");

    try 
    {           
      return TFP.ReadFields();
    }
    catch (MalformedLineException)
    {
      StringBuilder m_sbLine = new StringBuilder();

      for (int i = 0; i < TFP.ErrorLine.Length; i++)
      {
        if (i > 0 && TFP.ErrorLine[i]== '"' &&(TFP.ErrorLine[i + 1] != ',' && TFP.ErrorLine[i - 1] != ','))
          m_sbLine.Append("\"\"");
        else
          m_sbLine.Append(TFP.ErrorLine[i]);
      }

      return parseCSVLine(m_sbLine.ToString());
    }
  }
}

A StreamReader is still used to read the CSV line by line, as follows:

using(StreamReader SR = new StreamReader(FileName))
{
  while (SR.Peek() >-1)
    myStringArray = parseCSVLine(SR.ReadLine());
}
RooiWillie
  • 2,198
  • 1
  • 30
  • 36
1

With Cinchoo ETL - an open source library, it can automatically handles columns values containing separators.

string csv = @"2,1016,7/31/2008 14:22,Geoff Dalgas,6/5/2011 22:21,http://stackoverflow.com,""Corvallis, OR"",7679,351,81,b437f461b3fd27387c5d8ab47a293d35,34";

using (var p = ChoCSVReader.LoadText(csv)
    )
{
    Console.WriteLine(p.Dump());
}

Output:

Key: Column1 [Type: String]
Value: 2
Key: Column2 [Type: String]
Value: 1016
Key: Column3 [Type: String]
Value: 7/31/2008 14:22
Key: Column4 [Type: String]
Value: Geoff Dalgas
Key: Column5 [Type: String]
Value: 6/5/2011 22:21
Key: Column6 [Type: String]
Value: http://stackoverflow.com
Key: Column7 [Type: String]
Value: Corvallis, OR
Key: Column8 [Type: String]
Value: 7679
Key: Column9 [Type: String]
Value: 351
Key: Column10 [Type: String]
Value: 81
Key: Column11 [Type: String]
Value: b437f461b3fd27387c5d8ab47a293d35
Key: Column12 [Type: String]
Value: 34

For more information, please visit codeproject article.

Hope it helps.

Cinchoo
  • 6,088
  • 2
  • 19
  • 34