LINQ for beginners

Question

Hello,

I love C#,I love the framework and I also love to learn as much as possible.Today I began to read articles about LINQ in C# and I couldn't find anything good for a beginner that never worked with SQL in his life.

I found this article very helpful and I understood small parts of it,but I'd like to get more examples.

After reading it couple of times,I tried to use LINQ in a function of mine,but I failed.

    private void Filter(string filename)
    {
        using (TextWriter writer = File.CreateText(Application.StartupPath + "\\temp\\test.txt"))
        {
            using(TextReader reader = File.OpenText(filename))
            {
                string line;
                while((line = reader.ReadLine()) != null)
                {
                    string[] items = line.Split('\t');
                    int myInteger = int.Parse(items[1]);
                    if (myInteger == 24809) writer.WriteLine(line); 
                }
            }
        }
    }

This is what I did and it did not work,the result was always false.

    private void Filter(string filename)
    {
        using (TextWriter writer = File.CreateText(Application.StartupPath + "\\temp\\test.txt"))
        {
            using(TextReader reader = File.OpenText(filename))
            {
                string line;
                while((line = reader.ReadLine()) != null)
                {
                    string[] items = line.Split('\t');
                    var Linqi = from item in items
                                where int.Parse(items[1]) == 24809
                                select true;
                    if (Linqi == true) writer.WriteLine(line); 
                }
            }
        }
    }

I'm asking for two things:

1. How would the function look like using as much Linq as possible?
2. A website/book/article about Linq,but please note I'm a decent beginner in sql/linq.

Thank you in advance!

Answer 1

For a website as a starting point, you can try Hooked on LINQ

Answer 2

101 LINQ Samples is certainly a good collection of examples. Also LINQPad might be a good way to play around with LINQ.

Answer 3

MSDN LINQ Examples: http://msdn.microsoft.com/en-us/vcsharp/aa336746.aspx

Answer 4

I got a lot out of the following sites when I started:

http://msdn.microsoft.com/en-us/library/bb425822.aspx

http://weblogs.asp.net/scottgu/archive/2007/05/19/using-linq-to-sql-part-1.aspx

Answer 5

cannot just check if Linqi is true...Linqi is an IEnumerable<bool> (in this case) so have to check like Linqi.First() == true

here is a small example:

string[] items = { "12121", "2222", "24809", "23445", "24809" };

                        var Linqi = from item in items
                                    where Convert.ToInt32(item) == 24809
                                    select true;
                        if (Linqi.First() == true) Console.WriteLine("Got a true");

You could also iterate over Linqi, and in my example there are 2 items in the collection.

Answer 6

If I was to rewrite your filter function using LINQ where possible, it'd look like this:

private void Filter(string filename)
{
    using (TextWriter writer = File.CreateText(Application.StartupPath + "\\temp\\test.txt"))
    {
        var lines = File.ReadAllLines(filename);
        var matches = from line in lines
                      let items = line.Split('\t')
                      let myInteger = int.Parse(items[1]);
                      where myInteger == 24809
                      select line;

        foreach (var match in matches)
        {
            writer.WriteLine(line)
        }
    }
}

Answer 7

To answer the first question, there frankly isn't too much reason to use LINQ the way you suggest in the above function except as an exercise. In fact, it probably just makes the function harder to read.

LINQ is more useful at operating on a collection than a single element, and I would use it in that way instead. So, here's my attempt at using as much LINQ as possible in the function (make no mention of efficiency and I don't suggest reading the whole file into memory like this):

private void Filter(string filename)
{
    using (TextWriter writer = File.CreateText(Application.StartupPath + "\\temp\\test.txt"))
    {
        using(TextReader reader = File.OpenText(filename))
        {
            List<string> lines;
            string line;
            while((line = reader.ReadLine()) != null)
                lines.Add(line);

            var query = from l in lines
                        let splitLine = l.Split('\t')
                        where int.Parse(splitLine.Skip(1).First()) == 24809
                        select l;

            foreach(var l in query)               
                writer.WriteLine(l); 
        }
    }
}

Answer 8

Well one thing that would make your sample more "LINQy" is an IEnumerable<string> for reading lines from a file. Here's a somewhat simplified version of my LineReader class from MiscUtil:

using System;
using System.Collections;
using System.Collections.Generic;
using System.IO;

public sealed class LineReader : IEnumerable<string>
{
    readonly Func<TextReader> dataSource;

    public LineReader(string filename)
        : this(() => File.OpenText(filename))
    {
    }

    public LineReader(Func<TextReader> dataSource)
    {
        this.dataSource = dataSource;
    }

    public IEnumerator<string> GetEnumerator()
    {
        using (TextReader reader = dataSource())
        {
            string line;
            while ((line = reader.ReadLine()) != null)
            {
                yield return line;
            }
        }
    }


    IEnumerator IEnumerable.GetEnumerator()
    {
        return GetEnumerator();
    }
}

Now you can use that:

    var query = from line in new LineReader(filename)
                let items = line.Split('\t')
                let myInteger int.Parse(items[1]);
                where myInteger == 24809
                select line;

    using (TextWriter writer = File.CreateText(Application.StartupPath 
                                               + "\\temp\\test.txt"))
    {
        foreach (string line in query)
        {
            writer.WriteLine(line);
        }
    }

Note that it would probably be more efficient to not have the let clauses:

    var query = from line in new LineReader(filename)
                where int.Parse(line.Split('\t')[1]) == 24809
                select line;

at which point you could reasonably do it all in "dot notation":

    var query = new LineReader(filename)
                        .Where(line => int.Parse(line.Split('\t')[1]) == 24809);

However, I far prefer the readability of the original query :)

Answer 9

First, I would introduce this method:

private IEnumerable<string> ReadLines(StreamReader reader)
{
    while(!reader.EndOfStream)
    {
        yield return reader.ReadLine();
    }
}

Then, I would refactor the main method to use it. I put both using statements above the same block, and also added a range check to ensure items[1] doesn't fail:

private void Filter(string fileName)
{
    using(var writer = File.CreateText(Application.StartupPath + "\\temp\\test.txt"))
    using(var reader = File.OpenText(filename))
    {
        var myIntegers =
            from line in ReadLines(reader)
            let items = line.Split('\t')
            where items.Length > 1
            let myInteger = Int32.Parse(items[1])
            where myInteger == 24809
            select myInteger;

        foreach(var myInteger in myIntegers)
        {
            writer.WriteLine(myInteger);
        }
    }
}

Answer 10

If you're after a book, I found LINQ in action from Manning Publications a good place to start.

Answer 11

As for Linq books, I would recommend:

  

Both are excellent books that drill into Linq in detail.

To add yet another variation to the as-much-linq-as-possible topic, here's my take:

using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;

namespace LinqDemo
{
    class Program
    {
        static void Main()
        {
            var baseDir = AppDomain.CurrentDomain.BaseDirectory;
            File.WriteAllLines(
                Path.Combine(baseDir, "out.txt"),
                File.ReadAllLines(Path.Combine(baseDir, "in.txt"))
                    .Select(line => new KeyValuePair<string, string[]>(line, line.Split(','))) // split each line into columns, also carry the original line forward
                    .Where(info => info.Value.Length > 1) // filter out lines that don't have 2nd column
                    .Select(info => new KeyValuePair<string, int>(info.Key, int.Parse(info.Value[1]))) // convert 2nd column to int, still carrying the original line forward
                    .Where(info => info.Value == 24809) // apply the filtering criteria
                    .Select(info => info.Key) // restore original lines
                    .ToArray());
        }
    }
}

Note that I changed your tab-delimited-columns to comma-delimited columns (easier to author in my editor that converts tabs to spaces ;-) ). When this program is run against an input file:

A1,2
B,24809,C
C

E
G,24809

The output will be:

B,24809,C
G,24809

You could improve memory requirements of this solution by replacing "File.ReadAllLines" and "File.WriteAllLines" with Jon Skeet's LineReader (and LineWriter in a similar vein, taking IEnumerable and writing each returned item to the output file as a new line). This would transform the solution above from "get all lines into memory as an array, filter them down, create another array in memory for result and write this result to output file" to "read lines from input file one by one, and if that line meets our criteria, write it to output file immediately" (pipeline approach).

Answer 12

I found this article to be extremely crucial to understand LINQ which is based upon so many new constructs brought in in .NET 3.0 & 3.5:

I'll warn you it's a long read, but if you really want to understand what Linq is and does I believe it is essential

http://blogs.msdn.com/ericwhite/pages/FP-Tutorial.aspx

Happy reading