tags:

views:

76

answers:

4

is there any way to read a csv file into a matrix, so every square in the file will be a cell in the matrix?

A: 

There are many open source CSV readers, and it's also easy to code your own.

For a start take look at codeplex.com: http://kbcsv.codeplex.com/

Or Codeproject tutorials: http://www.codeproject.com/KB/database/CsvReader.aspx

For sake of completion, here is my own utility class to read a line from a CSV file:

    /// <summary>
    /// Defines CSV reader states
    /// </summary>
    enum State
    {
        Initial, 
        Quote,
        Data,
        NestedQuote
    }

    /// <summary>
    /// Initializes a new instance of the <see cref="CsvReader"/> class.
    /// </summary>
    /// <param name="inputStream">The input stream.</param>
    public CsvReader(Stream inputStream)
    {
        if (inputStream == null) 
            throw new ArgumentNullException("inputStream");

        reader = new StreamReader(inputStream);
    }

    /// <summary>
    /// Reads a single line of CSV data.
    /// </summary>
    /// <returns>Array of CSV fields</returns>
    public string[] Read()
    {
        var line = reader.ReadLine();
        var retval = new List<string>();

        if (line == null) 
            return null;

        var state = State.Initial;
        var text = new StringBuilder();

        foreach (var ch in line)
            switch (state)
            {
                case State.Initial:
                    if (ch == '"') 
                        state = State.Quote;
                    else if (ch == ',') 
                        retval.Add(string.Empty);
                    else
                    {
                        text.Append(ch);
                        state = State.Data;
                    }

                    break;

                case State.Data:
                    if (ch == ',')
                    {
                        retval.Add(text.ToString());
                        text.Length = 0;
                        state = State.Initial;
                    }
                    else 
                        text.Append(ch);

                    break;

                case State.Quote:
                    if (ch == '"')
                        state = State.NestedQuote;
                    else 
                        text.Append(ch);

                    break;

                case State.NestedQuote:
                    if (ch == '"')
                    {
                        text.Append('"');
                        state = State.Quote;
                        break;
                    }

                    state = State.Data;
                    goto case State.Data;
            }

        retval.Add(text.ToString());

        return retval.ToArray();
    }

    /// <summary>
    /// Performs application-defined tasks associated with freeing, releasing, or resetting unmanaged resources.
    /// </summary>
    public void Dispose()
    {
        reader.Dispose();
    }

To make the matrix (untested):

var data = new List<string[]>();
string[] line;

using(reader = new CsvReader(stream))
  while((line = reader.Read()) != null)
    data.Add(line);

result = data.Select(row => row.Select(cell => int.Parse(cell)).ToArray()).ToArray();
sukru
Does that handle lines that contain carriage returns within quotes?
ck
this is what i want to do:int [][] mat=READ FROM CSV FILE...;my csv file contains int...
no, unfortunately carriage returns cause a new row to start : var line = reader.ReadLine(); -- I tried with Excel (2010), and it also did not allow line breaks in quotes, but I don't exactly know the specs
sukru
ok. but there are couple of things that i don't understand:where is "reader" defiend?2.is the function CsvReader is the constructor? and what is the parameter that it gets?3. where should i add the code you added now?
The whole code is the definition of a class - named CsvReader :). The reader is actually any input stream. You can have one by "var reader = new FileStream("path")"
sukru
ok, i made a couple of changes in your code to fit my needs, and it works great! tanks!
I'm glad that I was able to help.
sukru
A: 

There are many ways. Starting with a byte by byte reader. It depends on your csv file format (with/without header, line endings, " or '). I've written my own class.

A good reader to start with:

http://www.stellman-greene.com/CSVReader/

Andreas Rehm
+1  A: 

There is a text reader in the VisualBasic namespace that can be used in C# and handles even horrible CSV files very well:

TextFieldParser

Just add a reference to Microsoft.VisualBasic in your project.

ck
i added referenc, but i stil can't see "FileIO.TextFieldParser "..
ohh, now i see... but how do i use the textFieldParser?
cool i didn't know about this, thanks ck
tenfour
+1  A: 

CSV parsing with regular expressions: http://www.hotblue.com/article0000.aspx?a=0006

To expand the concept with custom separators, see this post: http://stackoverflow.com/questions/3147995/regex-matching-notcaracter-string-c-regex

Peet Brits