Comparing two hashsets

Question

I have two hashsets that are loading data from two different text files. The contents of both text files look as follows:

name/12441431252132
name1/323244231244142
name2/32423452524234

My code to currently load both files and make sure i only have unique results from textFile2:

HashSet<string> txt1 = new HashSet<string>(File.ReadLines("textFile1.txt"));
HashSet<string> txt2 = new HashSet<string>(File.ReadLines("textFile2.txt"));

txt2.ExceptWith(txt1);

My problem is that it only removes lines based on the criteria if the whole line matches. I want to remove It based on the name basis instead. For example, name2 should never be included if its in textFile1 even if the id's after the / are different.

How would i accomplish this?

Let me know if my explanation is not good i will try to improve it - and please excuse my english!

Not always there can be more than 1 of the same in textfile1, but if they are it should never be included after the ExceptWith in txt2 — hariette, May 22 '12 at 15:13

score 3 · Accepted Answer · answered May 22 '12 at 15:15

3

You can just add some string splitting to separate names and the rest of the content - the approach is a little "dirty" so in real code I would probably use foreach loops and introduce dedicated classes:

var content = File.ReadLines("textFile1.txt").Select(line => 
{
    var parts = line.Split('/');
    return new 
    { 
        Name = parts[0],
        Content = parts[1]
    };
});

HashSet<string> names = new HashSet<string>(content.Select(c=> c.Name));
HashSet<string> txt2 = new HashSet<string>(File.ReadLines("textFile2.txt"));
var uniques = txt2.Where(line => !names.Contains(line.Split('/')[0]));

answered May 22 '12 at 15:15

BrokenGlass

158,293
28
286
335

How would i get the first unique name and ID from the uniques var? later on in the code i need to use the first ones and make them into strings – hariette May 22 '12 at 15:22
@hariette: It wasn't clear to me that you wanted to separate out the lines into Name/Value pairs for `txt2` as well - in that case just use the same approach as used for the `contents` parsing – BrokenGlass May 22 '12 at 15:24

score 0 · Answer 2 · answered May 22 '12 at 15:20

if you split by /, you can build a HashSet of names that appear in the first set, then pick the items in the second set that have a name that does not appear in the first set.

var nameValues1=
    File
     .ReadLines(fileName)
     .Select(line=>line.Split('/'))
     .Select(parts=>new {name=parts[0],value=parts[1]});
var nameValues2=
    File
     .ReadLines(fileName2)
     .Select(line=>line.Split('/'))
     .Select(parts=>new {name=parts[0],value=parts[1]});
var names1=new HashSet<string>(nameValues1.Select(nv=>nv.name);
var result=
    nameValues2
     .Where(nv=>!names1.Contains(nv.name))
     .Select(nv=>string.Format("{0}/{1}",nv.name,nv.value);

Tim Schmelter · Answer 3 · 2012-05-22T15:44:29.233

0

Are you sure that a HashSet is still the best choice? Here is a different approach using a Dictionary<String, String>:

var lines1 = System.IO.File.ReadLines(path1);
var lines2 = System.IO.File.ReadLines(path2);
var allItems = new Dictionary<String, String>();
foreach (var line in lines1.Concat(lines2))
{
    String[] tokens = line.Split('/');
    if (tokens.Length == 2)
    {
        String name = tokens[0];
        String number = tokens[1];
        if (!allItems.ContainsKey(name))
            allItems.Add(name, number);
    }
}

edited May 22 '12 at 15:44

answered May 22 '12 at 15:30

Tim Schmelter

450,073
74
686
939

That's problematic, given the OP states that names in a file are **not** unique. – spender May 22 '12 at 15:31
@spender: But his requirement is to filter out the duplicate names. Above approach would override existing names' numbers, but could be changed easily to do something else. – Tim Schmelter May 22 '12 at 15:32
1

@spender: Since OP just want to take the first, i've edited above to check for existence with `ContainsKey`. – Tim Schmelter May 22 '12 at 15:45

Comparing two hashsets

3 Answers3

Linked