How to speed up Word Interop processing?

Question

I am Very new at C# and have written a fairly clunky code. I have been doing a lot of courses online and a lot say that there are several ways to approach problems. Now i have m

Answer 1

Ok Completed it so we now process all information as before and import the whole document still. Total runtime is 02:09.8 for 2780 sentences and approx 44,000 words (that includes spaces!) Below is my (Not perfect Code) not bad considering I picked up C# 2 weeks ago ;) Hope this helps someone out in the future.

    public Form1()
    {
        InitializeComponent();
    }
    public int id = 0;
    public int[] iD = new int[400];
    public string[] timeOn = new string[400];
    public string[] timeOff = new string[400];
    public string[] dutyNo = new string[400];
    public string[] day = new string[400];
    public string[] hours = new string[400];

    //Create File Location Var
    public string fileLocation = null;

    // On Click of Add Dutys
    private void button1_Click(object sender, EventArgs e)
    {
        //Sets Progress Bar visible and prepares to increment
        pBar1.Visible = true;
        pBar1.Minimum = 1;
        pBar1.Value = 1;
        pBar1.Step = 1;


        //Stopwatch test Declared
        Stopwatch stopWatch = new Stopwatch();

        try {
            //Self Test to see if a File Location has been set for Duty document.
            if (fileLocation == null) {
                //If not set prompts user with message box and brings up file explorer
                MessageBox.Show("It Appears that a file location has not yet been set, Please Select one now.");
                Stream myStream = null;
                OpenFileDialog openFileDialog1 = new OpenFileDialog();
                //Sets default Location and Default File type as .doc
                openFileDialog1.InitialDirectory = "c:\\";
                openFileDialog1.Filter = "All files (*.*)|*.*|Word Files (*.doc)|*.doc";
                openFileDialog1.FilterIndex = 2;
                openFileDialog1.RestoreDirectory = true;
                //Waits for User to Click ok in File explorer and then Sets file location to var
                if (openFileDialog1.ShowDialog() == DialogResult.OK)
                {
                    try
                    {
                        //Checks to make sure a file location is set
                        if ((myStream = openFileDialog1.OpenFile()) != null)
                        {
                            using (myStream)
                            {
                                //This is where we set location to var
                                fileLocation = openFileDialog1.FileName;
                            }
                            //Prompts user to click a file before OK
                        }else { MessageBox.Show("Please Select a file location before clicking ok"); }
                    }
                    catch (Exception ex)
                    {
                        MessageBox.Show("Error: Could not read file from disk: " + ex.Message);
                    }
                }
            }

           //Loads New Duty file 
            Microsoft.Office.Interop.Word.Application application = new Microsoft.Office.Interop.Word.Application();
            Microsoft.Office.Interop.Word.Document document = application.Documents.Open(fileLocation);
            //Begin stop watch (COPY TIME)
            stopWatch.Start();

            //Sets Count to No of sentences and then prepares Array using Number of sentences 
            //**This process reduces amount of processng time by taking everything in to the program to start and then dealing with it.
            int count = document.Sentences.Count;
            string[] sents = new string[count];
            //Then sets the Progress bar to the Number of sentences that will be Copied to our array
            pBar1.Maximum = count;

            try {
                //For loop runs throug every sentence and adds it to the array.
                for (int i = 0; i < count; i++) {
                    sents[i] = document.Sentences[i+1].Text;
                    //increment Progress bar by 1 for every sentence(Parse made)
                    pBar1.PerformStep();
                }
                //Closes our instance of word
                application.Quit();
                try {

                    for (int i = 0; i < count; i++)
                    {
                        //Sets our Split criteria 
                        char[] delimiterChars = { ' ','\t' };
                        string[] test = (sents[i].Split(delimiterChars));
                        //we then enter For loop that runs for the number of ords found/Split
                        for (int a = 0; a < test.Length; a++)
                        {  
                            //If tests only begin if the word is NOT a space blank, tab , - As these do parse through into our Test arrays
                            if (!(test[a] == "" || test[a].Contains("/t")|| test[a].Contains("-") || test[a].Contains(" ")))
                            {
                                //If tests to find Duty numbers ours on off and assigns ID number for easy indexing. 
                                //##THIS DOES ASSUME THAT OUR TIMES ARE 1 SPACE AFTER THEIR IDENTIFIERS.
                                if (test[a] == "TG")
                                {
                                    dutyNo[id] = test[a + 2]; 
                                }
                                else if (test[a] == "On")
                                {
                                    iD[id] = id;
                                    timeOn[id] = test[a + 1];
                                }
                                else if (test[a] == "Off")
                                {
                                    timeOff[id] = test[a + 1];
                                }
                                else if (test[a] == "Hrs")
                                {
                                    hours[id] = test[a + 1];
                                }
                                else if (test[a] == "Days")
                                {
                                    day[id] = test[a + 1];
                                    //PRINTS TO USER VIA LIST BOX ALL THE DUTYS ADDED.
                                    listBox1.Items.Add(string.Format("ADDED:Duty No:{3} Time on:{1} Time off:{2} Hours{5} Day:{4} ID:{0}", iD[id], timeOn[id], timeOff[id], dutyNo[id], day[id], hours[id]));
                                    id++;
                                }

                            }
                        }
                    }
                }
                catch(Exception ex) { MessageBox.Show("Error in split:" + ex.Message); }
            }
            catch(Exception ex) { MessageBox.Show("error setting string to Document:" + ex.Message); }
            //Stopwatch Is then printed for testing purposes.
            TimeSpan ts = stopWatch.Elapsed;
            string elapsedTime = String.Format("{0:00}:{1:00}:{2:00}.{3:00}", ts.Hours, ts.Minutes, ts.Seconds,
            ts.Milliseconds / 10);
            Console.WriteLine("RunTime (total):" + elapsedTime);

            stopWatch.Reset();

        }
        catch(Exception ex) { MessageBox.Show("Error in reading/finding file: "+ ex.Message); }

    }


}

}

I use all this code with a fairly large List box (ListBox1) a Button (Button1)and a Non visible on start-up Progress bar (pBar1).

Answer 2

You can load the entire .Content range with OpenXml and process it like that, then reimport it

Answer 3

Office Interop is fairly slow.

Openxml may have been faster, but the file is .doc, so it probably won't be able to handle it.

---

But just like with Excel in this question there is a way you can improve the performance - do not access each word in a Range by index, because AFAIK it causes creation of a separate Range instance wrapped in RCW, and that is primary candidate for a performance bottleneck in your application.

That means that your best bet to improve the performance is to load all the words (.Text) into some indexable collection of Strings before the actual processing, and only then use that collection to create the output.

How to do it in the fastest way? I am not exactly sure, but you can try either getting all the words from _Document.Words enumerator (though it may or may not be more performant, but at least you will be able to see how long it takes to just retrieve the required words):

var words = document
    .Cast<Range>()
    .Select(r => 
        r.Text)
    .ToList();

or you may try to use _Document.Content range Text, though you would then have to separate individual words by yourself.

Answer 4

Instead of using:

document.Words[i].Text

multiple times, do:

String Text = document.Words[i].Text;

at the top of the for loop and use "Text" (or whatever you want to call it) instead. Eugene Podskal's suggestions seem very helpful but this simple improvement (that I was thinking of before seeing Eugene's response) is very easy to do and could make a substantial improvement.