BackgroundWorker Question (C# Windows Forms)

I am on MSDN reading about the BackgroundWorker class and I have a question about how it works.

The following code has a for loop in it. And inside the for loop, in the else clause, you're supposed to:

Perform a time consuming operation and report progress.

But, why is there a for loop, and why is its开发者_StackOverflow社区 maximum value only 10?

private void bw_DoWork(object sender, DoWorkEventArgs e)
{
    BackgroundWorker worker = sender as BackgroundWorker;

    for (int i = 1; (i <= 10); i++)
    {
        if ((worker.CancellationPending == true))
        {
            e.Cancel = true;
            break;
        }
        else
        {
            // Perform a time consuming operation and report progress.
            System.Threading.Thread.Sleep(500);
            worker.ReportProgress((i * 10));
        }
    }
}

I have a really massive database, and it takes sometimes up to a minute to check for new orders based on certain criteria. I don't want to guess how long it may take to complete a query, I want actual progress. How can I make this background worker report progress based on a MySQL SELECT query?

---

How can I make this background worker report progress based on a MySQL SELECT query?

You can't. That's one of the problems with a synchronous method call that you cannot predict ahead of time how long it is going to take. You have two cut points of time to deal with. Before you call the method, and after you call the method. You do not get anything in between. Either the method has returned, or it has not.

You can use statistics to your advantage though. You can record how long it takes each time it executes, store that, and use that to calculate a prediction, but it's never going to be accurate. With such a prediction, you could space out progress reporting accordingly so that you end up at 100% at or around the statistical prediction you've calculated.

However, if the database is slower or faster than usual, it'll be off.

Also note that whichever thread that is calling into MySQL to retrieve data can not be the same thread that is reporting progress, since it will be "waiting" for the MySQL database and the .NET code that talks to it to return with the data, all in one piece. You need to spin up yet another thread that reports the progress.

---

But, why is there a for loop, and why is its maximum value only 10?

In the example, the worker is reporting progress between 10 an 100, purely out of simplicity. The values 10 to 100 come from i (1-10), and the * 10 in ReportProgress.

The documentation says that ReportProgress takes:

The percentage, from 0 to 100, of the background operation that is complete.

When you write it for your really massive database, you must report progress as a percentage, between 0 and 100.

Given that your database may take "up to a minute", 1% is slightly more than 1/2 second, so you should see any associated progress bar move every 1/2 second or so. That sounds like pretty smooth reporting to me.

(Other answers describe why its difficult to attach the progress to a SQL-query)

---

You'll need to figure out a way to measure the progress of your query. Instead of one long query, you might be able to do it in batches (say of 10, then the progress increments by 10% each time).

---

The example is showing how to batch long processes so they can be reported. The 'sleep' instruction in the example would be replaced by a call to a method that did a time-consuming, batchable job.

In your case, unless you can split up your query into multiple parts, you can't really use ReportProgress to give feedback - you won't have any progress to report. A SQL query is a one-shot run, and ReportProgress is used for batchable things.

You may want to look into optimizing your database - it's possible that an index on a heavily-used table or something similar could be a big help. If this isn't possible, you'll have to find a way to do batched queries against the data (or get back the whole thing and go through it in code - ugh) if you want to be able to report meaningful progress.

---

The example code is just that: An example. So the 10 is arbitrary. It simply shows an example of estimating progress. In this case there are 10 discrete steps, so it can estimate progress easily.

I don't want to guess how long it may take to complete a query, I want actual progress.

A database query provides no means to report progress. You cannot do anything but guess.

What I do is this:

Assume that the longest time it will take is the timeout period for the connection. This way if the query fails because the connection died, the user will get a perfectly accurate progress bar. Most queries take far, far less time than the timeout value, so the user sees a little progress and then suddenly it completes. This gives the user the illusion that things happened better than expected!

To accomplish it I perform the Db query asyncronously and run the progress bar off a UI timer rather than using a BackgroundWorker.