Problem with LINQ

Question

I am converting PDF to TIFF and converting page by page to a temporary folder and at the end I will fetch those files to join them as in the original file. My code is as follows: (I used what is in my reply question):

string pastapdf = AppDomain.CurrentDomain.BaseDirectory + "temporario\\" + Path.GetFileNameWithoutExtension(e[i]);
Directory.CreateDirectory(pastapdf);
string pdfFile = Path.Combine(pastapdf, e[i]);
for (int h = 0; h <= (pdfReader.NumberOfPages - 1); h++)
{
    using (var document = PdfiumViewer.PdfDocument.Load(pdfFile))
    {
         var image = document.Render(h, dpi, dpi, true);
         image.Save(pastapdf + "\\" + Path.GetFileNameWithoutExtension(e[i]) + "_" + h.ToString() + ".tif", ImageFormat.Tiff);
         image.Dispose();
    }
}
string save = output + "\\" + Path.GetFileNameWithoutExtension(e[i]) + ".tif";
string[] tifffiles = GetFiles(pastapdf, "*.tif");
var finalfiles = tifffiles.OrderBy(f => 
    new string(f.Where(char.IsLetter).ToArray())).ThenBy(f => { 
    int number; if (int.TryParse(new string(f.Where(char.IsDigit).ToArray()), out number)) 
    return number; return -1; } ).ToList(); //código LINQ para ordenar

JoinTiffLZW(finalfiles.ToArray(), save);
Invoke(new Action(() => { pbar.PerformStep(); pdfReader.Dispose(); }));

You should do: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11.... And it’s doing: 1, 10, 11, 2, 3, 4, 5, 6, 7, 8, 9....

My question is if anyone can tell me another way to sort the list or what’s wrong with my code?

TEST

I uploaded a 40-page PDF. I divided and converted them all, then I fetched the files:

And then I should order with LINQ, but the result is this:

And it should be in the right order, so when I get everybody together, it’s the wrong order.

Answer 1

Probably this code has never worked before, at least not in the test case that was quoted in your question.

What happens in your code is this.

1. GetFiles returns a list of strings (in fact, it would be good to have mentioned this in the question).

   From the image, I saw that the list of strings would be something like

   var tifffiles = new[] 
   { 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_0", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_2", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_1", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_5", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_11", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_4", 
       "C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_10",  
   };
   

2. In this string list is called the method OrderBy with the following sorting function

   new string(f.Where(char.IsLetter).ToArray()
   

   This means that for each element of the collection (that of step 1) a new one will be created string containing only the letters - because only letters return true in the call for char.IsDigit.

   That is, the elements used in sorting will be - note that this does not alter the original elements:

   CQualquercoisavanualanual
   CQualquercoisavanualanual
   CQualquercoisavanualanual
   CQualquercoisavanualanual
   CQualquercoisavanualanual
   CQualquercoisavanualanual
   CQualquercoisavanualanual
   

   Therefore, the ordered collection that is returned keeps all elements in the same position that they were.

3. A call is made to the method ThenBy, using the following sorting function:

   int number; 
   if (int.TryParse(new string(f.Where(char.IsDigit).ToArray()), out number)) 
       return number; 
   return -1;
   

   What this function does - for each item in the collection - is the following:

   1. f.Where(char.IsDigit) selects all numbers from the original string;
   2. new string creates a new string with the selected numbers;
   3. int.TryParse tries to convert the generated string into a 32-bit integer (point of attention) and returns true if the conversion works and false otherwise; and
   4. return returns the converted value if the conversion worked or -1 otherwise.
      
      

   Now let’s try to apply this to the first element of array original

   C:\\Qualquercoisa_v2\\anual_2009\\anual_2009_0
   

   1. Select all numbers: 2200920090
   2. Create a string with these numbers: "2200920090"
   3. Try to convert to int: Problem here

   
   

   See the test code in repl.it

   Now just understand what is causing this error =)

   What happens there is that the numbers are larger than an integer of 32 bits. For this specific case (and several more) you can simply switch to a 64-bit integer (long). That is, change the second sort function to

   long number; 
   if (long.TryParse(new string(f.Where(char.IsDigit).ToArray()), out number)) 
   ... 
   

My advice is actually to review this code and try to rewrite it.

Answer 2

The idea here will be to create an object of anonymous type to assist sorting, with the filename and value to sort... this example takes into account that the value to sort is always between the last character _ and the .

    string[] files = new string[] {

        @"C:\Users\Utilizador\Pictures\revenche_2019_0.png",
        @"C:\Users\Utilizador\Pictures\revenche_2019_1.png",
        @"C:\Users\Utilizador\Pictures\revenche_2019_11.png",
        @"C:\Users\Utilizador\Pictures\revenche_2019_2.png",
        @"C:\Users\Utilizador\Pictures\revenche_2019_12.png",
        @"C:\Users\Utilizador\Pictures\revenche_2019_3.png"
    };

    var result = files.Select(f => new
    {
        file = f,
        index = int.Parse(string.Concat(f.Skip(f.LastIndexOf('_') + 1).TakeWhile(f2 => f2 != '.')))
    }).OrderBy(an => an.index)
    .Select(an => an.file)
    .ToList();

    result.ForEach(f => Console.WriteLine(f));