9
\$\begingroup\$

I am writing a new web service that will be performing a lot of large data load operations. To do so I am moving the data to a temporary table then merging the data in to the live data via a stored procedure.

That part I have down and working and am very confident in. What I am doing now is trying to optimize the loading code to help take the load off of the IIS server that will be performing these operations.

Here is what I am currently doing synchronously and works fine.

public void BulkLoadData(IDataReaderWithMappings reader, 
                         string createTempTableStatement, 
                         string loadingStoredProcedureName)
{
    using (var connection = new SqlConnection(_connectionString))
    using (var bulkCopy = new SqlBulkCopy(connection))
    using (var command = new SqlCommand(createTempTableStatement, connection))
    {
        connection.Open();
        command.ExecuteNonQuery();
        bulkCopy.DestinationTableName = "#temp";
        var mappings = reader.GetMappings();
        foreach (var sqlBulkCopyColumnMapping in mappings)
        {
            bulkCopy.ColumnMappings.Add(sqlBulkCopyColumnMapping);
        }
        bulkCopy.WriteToServer(reader);
        command.CommandType = CommandType.StoredProcedure;
        command.CommandText = loadingStoredProcedureName;
        command.ExecuteNonQuery();
    }
}

And here is my attempt at translating that in to asynchronous code.

public async Task BulkLoadDataAsync(IDataReaderWithMappings reader, 
                                    string createTempTableStatement, 
                                    string loadingStoredProcedureName)
{
    using (var connection = new SqlConnection(_connectionString))
    using (var bulkCopy = new SqlBulkCopy(connection))
    using (var command = new SqlCommand(createTempTableStatement))
    {
        var firstStage =  connection.OpenAsync()
                                    .ContinueWith(async delegate
        {
            await command.ExecuteNonQueryAsync().ConfigureAwait(false);
        });

        bulkCopy.DestinationTableName = "#temp";
        var mappings = reader.GetMappings();
        foreach (var sqlBulkCopyColumnMapping in mappings)
        {
            bulkCopy.ColumnMappings.Add(sqlBulkCopyColumnMapping);
        }

        //Wait for the connection opening and the create table task to finish.
        await firstStage.ConfigureAwait(false);

        await bulkCopy.WriteToServerAsync(reader).ConfigureAwait(false);
        command.CommandType = CommandType.StoredProcedure;
        command.CommandText = loadingStoredProcedureName;
        await command.ExecuteNonQueryAsync().ConfigureAwait(false);
    }
}

The part I am not sure of my self about is that .ContinueWith( section. I thought it would be good to start opening the connection and creating the temporary table while I populate the bulk copy column mappings, and I have used async/await many times before but this is the first time I have used .ContinueWith( in conjunction with the async/await syntax.

Did I do it correctly, or should that await firstStage; be await firstStage.Unwrap();? I still do not exactly understand when I should be using Unwrap(). EDIT: Ok, Unwrap() is not listed so I could not use it even if I wanted to, but if anyone wants to include a explanation in their answer of when I should I would appreciate it.

(P.S. Do not worry about SQL injection from createTempTableStatement the strings being passed in are hard-coded resource strings with no user input.)


With the suggestions from svick and some other tweaks I made while waiting for an answer here is the final version

public async Task BulkLoadDataAsync(IDataReader reader, 
                                    string createTempTableStatement, 
                                    string loadingStoredProcedureName, 
                                    params SqlParameter[] args)
{
    using (var connection = new SqlConnection(_connectionString))
    using (var bulkCopy = new SqlBulkCopy(connection))
    using (var command = new SqlCommand(createTempTableStatement))
    {
        Func<Task> firstStageFunc = async () => 
        {
            await connection.OpenAsync().ConfigureAwait(false);
            await command.ExecuteNonQueryAsync().ConfigureAwait(false);
        };

        var firstStage = firstStageFunc();

        bulkCopy.DestinationTableName = "#temp";
        var mappedReader = reader as IDataReaderWithMappings;
        if (mappedReader != null)
        {
            var mappings = mappedReader.GetMappings();
            foreach (var sqlBulkCopyColumnMapping in mappings)
            {
                bulkCopy.ColumnMappings.Add(sqlBulkCopyColumnMapping);
            }
        }

        //Wait for the connection opening and the create table task to finish.
        await firstStage.ConfigureAwait(false);
        var bulkCopyTask = bulkCopy.WriteToServerAsync(reader);

        command.CommandType = CommandType.StoredProcedure;
        command.CommandText = loadingStoredProcedureName;

        if (args != null)
            command.Parameters.AddRange(args);

        //Wait for the bulk copy to finish.
        await bulkCopyTask.ConfigureAwait(false);
        await command.ExecuteNonQueryAsync().ConfigureAwait(false);
    }
}
\$\endgroup\$

1 Answer 1

3
\$\begingroup\$

If you want to combine ContinueWith() and async this way, then you need to use Unwrap(). (And it is available, though my ReSharper doesn't seem to think so, maybe you use that too?) If you don't do that, you're not actually waiting for ExecuteNonQueryAsync() to complete.

But using ContinueWith() is usually not a good idea when you can use await. Instead, you can use a helper method:

private static async Task FirstStageAsync(SqlConnection connection, SqlCommand command)
{
    await connection.OpenAsync().ConfigureAwait(false);
    await command.ExecuteNonQueryAsync().ConfigureAwait(false);
}

And then in the main method:

var firstStage = FirstStageAsync(connection, command);

…

//Wait for the connection opening and the create table task to finish.
await firstStage.ConfigureAwait(false);

Another option would be use a lambda instead of the helper method:

Func<Task> firstStageFunc = async () =>
{
    await connection.OpenAsync().ConfigureAwait(false);
    await command.ExecuteNonQueryAsync().ConfigureAwait(false);
};

var firstStage = firstStageFunc();
\$\endgroup\$
1
  • \$\begingroup\$ I do use resharper, and actually I ended up re-writing the code to use a helper method anyway since I asked the question because I thought it looked "cleaner". \$\endgroup\$ Mar 17, 2014 at 17:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.