Tuesday 23 November 2010

Drupal 7 and SQL Server Spatial Data

In a previous post, I explored integration between Drupal 7 and SQL Server Reporting Services. In this post, I’ll examine how you can take advantage of the fact that Drupal 7 can be installed on SQL Server to store geographic data using SQL Server 2008’s spatial data support, and also how you can integrate Bing Maps and Twitter geographic data into your Drupal site.

Drupal and SQL Server Spatial Data Demo

Microsoft introduced support for spatial data in SQL Server 2008 (see some of my past posts for details), and given that you can now install Drupal on Windows with a SQL Server content database, it makes sense to use this support to extend your Drupal site to include content with a geographical element. The scenario we decided to use when exploring this idea was a site for a local arts festival that includes the ability to publish details of events, including their location. However, you could easily adapt the idea for any content that has a location-based element. The video shows the solution we created (you can also view the video on YouTube).

Our solution not only incorporates spatial data that is stored in SQL Server, but also uses the Bing Maps control to visualize the spatial data, and retrieves tweets that include a geographic location from Twitter. This sort of “mash-up” approach where data and services from multiple sources on the Web are combined to create a rich user experience is the very essence of today’s most cutting edge Web applications, and when combined with the social/community-based publishing capabilities of Drupal it can be used to create some pretty compelling compelling solutions.

The diagram below shows the high-level architecture of the solution. Note that the SQL Server database forms the “data hub” of the application, storing the geographical data from both the Drupal content and the related tweets (which are retrieved and cached in the database periodically using the Cron scheduler in Drupal).

SpatialArchitecture

At the heart of the solution is a set of custom Drupal modules that we implemented in PHP. These are packages and installed in the Drupal site to add the required spatial data functionality. The first module we created, and the one on which the others all depend, is a module named GeoField that defines a Location field type that can be included in Drupal content types. The Location field type stores a geographical Longitude-latitude point in well-known text (WKT) format, which can easily be instantiated as a SQL Server geography data type instance. The module also defines a widget for setting Location values, which consists of a textbox and a Bing Maps control. Users can enter an address in the text box and geocode it using the Bing Maps API, or they can right-click a point on the map and reverse-geocode the address. The code to do the geocoding is implemented in JavaScript as described in a previous blog post. The following PHP functions define the field info and schema within Drupal:

function geo_field_field_info() {
  return array(
    'geo_field_Location' => array(
      'label' => t('Location Data'),
      'description' => t('A field for storing geographic data.'),
      'default_widget' => 'geo_field_bingmap',
      'default_formatter' => 'geo_field_simple_text',
    ),
  );
}

function geo_field_field_schema($field) {
  $columns = array(
    'location' => array('type' => 'varchar', 'length' => 255, 'not null' => FALSE),
  );
  $indexes = array(
    'location' => array('location'),
  );
  return array(
    'columns' => $columns,
    'indexes' => $indexes,
  );
}

The Bing Maps widget, and an alternative simple textbox-based widget, is defined with the following PHP code:

function geo_field_field_widget_form(&$form, &$form_state, $field, $instance, $langcode, $items, $delta, $element) {
   $value = isset($items[$delta]['location']) ? $items[$delta]['location'] : '';
  $element += array(
    '#delta' => $delta,
  );
  $element['location'] = array();

  switch ($instance['widget']['type']) {

    case 'geo_field_bingmap':
    drupal_add_js('
http://dev.virtualearth.net/mapcontrol/mapcontrol.ashx?v=6.3', 'external');
      $element['location'] += array(
        '#prefix' => '',
        '#attributes' => array('class' => array('edit-field-geomap')),
        '#attached' => array(
       
         'js' => array(drupal_get_path('module', 'geo_field') . '/BingMaps.js'),       
        ),
        '#suffix' => '<div class="bing-map-control form-item form-type-textfield"><label for="txtAddress">Address</label><input type="text" id="txtAddress" class="form-text bing-map-address" size="60"/><input type="button" value="Find" class="location-search-button form-submit" class="form-submit" /><br/><div id="mapDiv" class="bing-map-edit bing-map" style="position:relative; width:400px; height:300px;"></div></div>',
      );
     

    case 'geo_field_geotext':
      $element['location'] += array(
       '#prefix' => '<div class="geo_field_geotext" style=" margin-top:20px; border: thin solid #808080;"><p>Enter a location:</p>',
        '#type' => 'textfield',
        '#default_value' => $value,
        '#size' => 50,
        '#maxlength' => 255,
        '#suffix' => '</div>',
      );
      break;

     }
  return $element;
}

Note the reference to the Bing Maps API and to the BingMaps.js JavaScript file, which includes the code used to handle the user interaction with the Bing Maps control.

The next task was to create a GeoEvent module that defines a content type for a festival event. This content type includes a Location field as well as more standard Drupal fields, such as a title, image, description, tags, and a simple text field to specify the date and time of the event.

The final module to be created was the GeoMap module, which defines the Event Map page on which events and tweets are plotted. The code in this module includes the following function, which retrieves the events from the Drupal database and returns them as a GeoRSS feed that can be imported into the Bing Maps control by client-side JavaScript code.

function geo_map_events() {
    global $base_url, $language_content;
  
    $results = db_query(
        'SELECT r.title, b.body_value what, r.nid, w.[event_location_location] loc, geography::STPointFromText(w.[event_location_location], 4326).AsGml() [where] '.
        'FROM node_revision r '.
        'LEFT JOIN field_revision_body b ON r.nid = b.revision_id '.
        'JOIN field_revision_event_location w ON w.revision_id = r.nid '.
        'JOIN node n ON r.nid = n.nid '.
        'WHERE n.type = \'event\'');
   
    $items = '';
    $addedItems = 0;
    foreach ($results as $result) {
        // We have to add in the namespaces here else the Bing control
        // can't parse the content.
        $xml = str_replace('<', '<gml:', $result->where);
        $xml = str_replace('gml:/', '/gml:', $xml);
       
        $description = $result->what != '' ? $result->what . '&lt;br /&gt;' : '';
        $description .= '&lt;a href=\'index.php?q=node/' . $result->nid . '\'&gt;More info&lt;/a&gt; &lt;br /&gt;';
        $description .= '&lt;a href="#' . $result->loc . '" class="twitter-button" &gt; &lt;img src="modules/geo_field/images/twitter.png" alt="Show nearby tweets" /&gt; &lt;/a&gt;';
       
        $item_text = '<entry>'.
                    '<title>' . $result->title . '</title>'.
                    '<description>' . $description .'</description>'.
                    '<georss:where>' . $xml .
                    '</georss:where>'.
                  '</entry>';

        $items .= $item_text;
        $addedItems++;
    }

  if ($addedItems == 0) {
    $placeholder_text  ='<entry>' .
      '<title>Placeholder</title>'.
      '<description></description>'.
      '<georss:where><gml:Point xmlns="http://www.opengis.net/gml"><gml:pos>-90 0</gml:pos></gml:Point>'.
            '</georss:where>'.
            '</entry>';
    $items .= $placeholder_text;
  }
 
  $output  = "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n";
  $output .= "<feed xmlns='http://www.w3.org/2005/Atom' xmlns:georss='http://www.georss.org/georss' xmlns:gml='http://www.opengis.net/gml'>\n";
  $output .=
        "<title>Festival Events</title>\n".
              "<subtitle>Events for the 2010 festival</subtitle>\n".
              "<author>\n".
                "<name>Drupal Administrator</name>\n".
              "</author>";
  $output .= $items;
  $output .= "</feed>\n";

  drupal_add_http_header('Content-Type', 'text/xml; charset=utf-8');
  print $output;
}

Note that the query used to retrieve the data uses the SQL Server geography type’s STPointFromText function to instantiate a geography instance from the WKT representation of each location, and the AsGml method to then convert that geography instance to Geographic Markup Language (GML) format. This enables the code to simply embed the GML representation of the location into the GeoRSS feed (since a GeoRSS feed is in effect an XML representation of a collection of GML entities). If no locations are found, a placeholder location is created to avoid returning an empty feed.

The client-side JavaScript to request this feed and display the results as a layer on the Bing Maps control is in a  script file named EventMaps.js, which is referenced by the Drupal page:

function showEvents() {
  eventLayer.DeleteAllShapes();
  var url = "index.php?q=geo_field/items_georss";
  var veLayerSpec = new VEShapeSourceSpecification(VEDataType.GeoRSS, url, eventLayer);
  map.ImportShapeLayerData(veLayerSpec, function(feed) {
        var numShapes = feed.GetShapeCount();
        for (c = 0; c < numShapes; c++) {
            var shape = feed.GetShapeByIndex(c);
     
      if (shape.GetDescription() == 'Placeholder' &&
          shape.GetTitle() == 'Placeholder') {
        eventLayer.DeleteAllShapes();
      }
     
            shape.SetCustomIcon("modules/geo_field/images/pin.png");
        }

    }, false);
   
    map.SetCenterAndZoom(new VELatLong(55.67630968987942, 12.569355145096778), 10);
}

Note that if the layer contains only a placeholder location (because there were no events), the location is deleted to produce an empty map.

The GeoMap module also includes a function that queries Twitter to return any tweets containing a specified hash-tag (which the administrator can set in the settings for the module). These tweets are then inserted into a table in the SQL Server database. The table includes a geography column in which the location of the tweet is stored.

function tweets_download() {
    $tweetsfeed = "http://search.twitter.com/search.json?q=%23" . variable_get('geomap_hashtag', 'drupalcon');
   
    $results = json_decode(file_get_contents($tweetsfeed), TRUE);
   
  // Insert items into the table if they're not already there.
    $insertquery =
        "IF NOT EXISTS ".
        "(SELECT * FROM dbo.tweets WHERE TwitterID = :tweetid1) ".
        "INSERT INTO dbo.Tweets VALUES " .
        "(:datetime, :tweeter, :content, geography::GeomFromGml(:location, 4326), :tweetid2)";
   
    foreach ($results['results'] as $tweet) {
        if ($tweet['geo'] != null) {
            $pointdetails =
                "<Point xmlns='http://www.opengis.net/gml'><pos>" .
                $tweet['geo']['coordinates'][0] . ' ' . $tweet['geo']['coordinates'][1] .
                "</pos></Point>";
               
            $timestamp = strtotime($tweet['created_at']);
            $sqldate = date('c', $timestamp);
           
            $sqldate = substr($sqldate, 0, strlen($sqldate) - 6);
       
            $params = array(
                    'tweetid1' => $tweet['id'],
                    'tweetid2' => $tweet['id'],
                    'datetime' => $sqldate,
                    'tweeter' => $tweet['from_user'],
                    'content' => $tweet['text'],
                    'location' => $pointdetails);
                           
            db_query(
                $insertquery,
                $params);
        }
    }
}

Note that the Transact-SQL query executed by the code uses the SQL Server geography data type’s GeomFromGml function to convert the value retrieved from Twitter into a geography instance before storing it in the table. This function is called by the Drupal Cron scheduler, so the table of tweets is updated periodically.

When the user clicks the Twitter icon for an individual event, the following PHP function is used to retrieve the tweets within 500m of the event and return them as a GeoRSS feed to be displayed on the map.

function geo_map_tweets_feed($location) {
  // Use the location functionality in MSSql to query all tweets near a given
  // location.
    $results = db_query(
        'SELECT    Tweeter + \' (\' + cast(TweetDateTime as nvarchar(50)) + \')\' as tweeter, '.
        'Tweet as tweet, '.
        'TweetLocation.AsGml() as [where] '.
        'FROM tweets '.
        'WHERE TweetLocation.STIntersects(geography::STPointFromText(\'POINT(' . $location . ')\', 4326).STBuffer(500)) = 1');
   
    $items = '';
    $found = false;
   
    foreach ($results as $result) {
        $found = true;
        $xml = str_replace('<', '<gml:', $result->where);
        $xml = str_replace('gml:/', '/gml:', $xml);

        $item_text = '<entry>'.
                    '<title>' . $result->tweeter . '</title>'.
                    '<description>' . $result->tweet .'</description>'.
                    '<georss:where>' . $xml .
                    '</georss:where>'.
                  '</entry>';

        $items .= $item_text;
    }
       
    if (!$found) {   
        $items = '<entry>'.
                    '<title>Placeholder</title>'.
                    '<description>Placeholder</description>'.
                    '<georss:where><gml:Point xmlns="http://www.opengis.net/gml"><gml:pos>-90 0</gml:pos></gml:Point>'.
                    '</georss:where>'.
                  '</entry>';
    }

  $output  = "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n";
  $output .= "<feed xmlns='http://www.w3.org/2005/Atom' xmlns:georss='http://www.georss.org/georss' xmlns:gml='http://www.opengis.net/gml'>\n";
  $output .= "<title>Festival Events</title>\n".
             "<subtitle>Events for the 2010 festival</subtitle>\n".
              "<author>\n".
                "<name>Drupal Administrator</name>\n".
              "</author>";
  $output .= $items;
  $output .= "</feed>\n";

  drupal_add_http_header('Content-Type', 'text/xml; charset=utf-8');
  print $output;   
}

Note that the WHERE clause of the Transact-SQL query executed by the code uses the STIntersects and STBuffer methods of the geography data type to find tweets within a 500m buffer of the event (which is instantiated as a geography instance by using the STPointFromText method as before).

As with the Reporting Services demo I blogged about previously, this solution we built was based on pre-beta releases of Drupal 7 and the related SQL Server driver for PDO and SQL Server module. To avoid the inevitable versioning issues that tend to arise with pre-release code dependencies, I haven’t includes the full source code. The plan is to publish it on CodePlex when the final release of Drupal 7 and the associated SQL Server components is available. Until then, hopefully you’ll find this article useful as an insight into the kinds of spatial data solutions you can build with Drupal and SQL Server.

Drupal 7 and SQL Server Reporting Services

Drupal is a popular content management system that until relatively recently has been associated more or less exclusively with the LAMP (Linux, Apache, MySQL, and PHP) Web development stack. However, Acquia Drupal is now available for Windows and can be installed with the Microsoft Web Platform Installer, and at DrupalCon in San Francisco, it was announced that a new SQL Server PHP driver for PDO and a SQL Server module for Drupal 7 would be released, enabling Drupal site data to be hosted in SQL Server. You can download a pre-release version of a Drupal and SQL Server package from here.

Drupal and Reporting Services Demo

Drupal is highly extensible through its PHP-based modular architecture, and SQL Server includes a wealth of features that could potentially enhance the value of a community-based content site, so given that you can now install Drupal with SQL Server, it seemed logical to do some investigation into how you can integrate some of the capabilities of SQL Server into a Drupal site.

Perhaps the most obvious starting point is to integrate Reporting Services into a Drupal site to create a sort of “Community Dashboard” that includes reports showing content and author related activity within the site. SQL Server Express Edition with Advanced Services includes Reporting Services for free, and the SQL Server Reporting Services SDK for PHP provides a framework for accessing Reporting Services from PHP-based applications. The video shows the Drupal solution we created using these technologies (you can also view this video on YouTube).

The demo actually shows “bi-directional” integration between Drupal and SQL Server Reporting Services in the sense that not only does the site include a custom PHP module to “pull” the reports from Reporting Services into the Drupal site user interface, the reports themselves query data in the underlying Drupal content database, which in this instance is hosted in SQL Server 2008 R2 Express Edition.

The diagram below shows the high-level architecture of the integration solution we created. Note that the Reporting Services instance contains two folders – one containing reports that show activity for all users, and another containing reports that are filtered based on the current Drupal user’s ID. The custom module in the Drupal site ascertains whether the current user is an administrator, and if not retrieves the current user’s ID via the Drupal PHP API and passes it as a hidden parameter when retrieving the reports.

Reporting Application Architecture

The reports query the tables in the Drupal content database. For example, the following query is used to retrieve the data for the Content Activity report:

SELECT n.Title Page,
       cast(dateadd(ss,n.created,'1/1/1970') as Date) CreatedDate,
       u.name Author,
       c.[Subject] Comment,
       uc.name [Commenter],
       cast(dateadd(ss,c.created,'1/1/1970') as Date) CommentDate,
       cn.totalcount PageViews
FROM dbo.Node n
JOIN dbo.Users u ON n.[uid] = u.[uid]
LEFT JOIN dbo.Comment c ON n.nid = c.nid
LEFT JOIN dbo.Users uc ON c.[uid] = uc.[uid]
LEFT JOIN node_counter cn ON n.nid = cn.nid
WHERE MONTH(cast(dateadd(ss,n.created,'1/1/1970') as Date)) = @Month
AND YEAR(cast(dateadd(ss,n.created,'1/1/1970') as Date)) = CONVERT(char(4),@Year)

Note that the date values in the Drupal database are stored in the standard UNIX format (the number of seconds since 1/1/1970), and so need to be cast to an appropriate SQL Server data type (we’ve used the Date datatype).

Our custom module code uses the SQL Server Reporting Services SDK for PHP to connect to and retrieve information from Reporting Services. For example, the following code is used to retrieve a list of available reports in the appropriate folder in Reporting Services (depending on whether or not the user is an administrator):

/*
  connect to SQL Server Reporting Services and retrieve a list of available reports
  */
function getReports(){
module_load_include('php', 'ssreport', 'bin/SSRSReport');
    
    $reports = array();
try
{
    $ssrs_report =  new SSRSReport(new Credentials(variable_get('ssreport_user'), variable_get('ssreport_password')), variable_get('ssreport_url'));
    $childPath = "";
 
// if user is admin then load from community dashboard
if (user_access('Administer site configuration')){
   $childPath = variable_get('ssreport_primary');
}
else{
// load from personal dashboard
   $childPath =  variable_get('ssreport_secondary');
}

$catalogItems = $ssrs_report->ListChildren($childPath, true);

foreach ($catalogItems as $catalogItem) {

if ($catalogItem->Type == ItemTypeEnum::$Report) {
$reports[$catalogItem->Name] = $catalogItem->Name;


}
}

}catch(SSRSReportException $serviceException)
{
    drupal_set_message($serviceException);
}
return $reports;

}

Note that the code uses a number of variable_get calls to retrieve global variables that are set in the Drupal administrative (such as the credentials to access the Reporting Services instance, the URL of the Reporting Services Web service, and the path to the CommunityDashboard and PersonalDashboard folders). Finally, the code uses the SSRSReport.ListChildren method provided by the SQL Server Reporting Services SDK for PHP API to retrieve the list of reports.

When a user selects a report, the following PHP function is used to retrieve the list of parameters that the report expects:

$rep=NULL;
if (user_access('Administer site configuration')){

$rep = t(variable_get('ssreport_primary').'/@report',array('@report' => $report));
 
}
else{
$rep = t(variable_get('ssreport_secondary').'/@report',array('@report' => $report));

}

// connect to SQL Server Reporting Services using the credentials and URL supplied in the administrative dashboard
$rs = new SSRSReport(new Credentials(variable_get('ssreport_user'), variable_get('ssreport_password')),variable_get('ssreport_url'));
$params = array();
  try
{
// get the parameters for the specified report
$reportParameters = $rs ->GetReportParameters($rep, null, true, null,     null);
    
$params = array();
        
foreach($reportParameters as $reportParameter)
            
{
    array_push($params,$reportParameter->Name);            
   
}

Finally, the following code is used to render the selected report in the Drupal user interface:

$rs = new SSRSReport(new Credentials(variable_get('ssreport_user'), variable_get('ssreport_password')),variable_get('ssreport_url'));
//get the report params array
  $path = drupal_get_path('module','report_services');

$reportParameters = $rs ->GetReportParameters($rep, null, true, null,     null);
    
$params = array();
        
foreach($reportParameters as $reportParameter)
{
  //code to set the parameter values

}

       $rs->SetExecutionParameters2($parameters);
   

    // delete existing image files
    $dirname = './images/';
   $dir_handle = opendir($dirname);
   if (!$dir_handle)
      return false;
   while($file = readdir($dir_handle)) {
      if ($file != "." && $file != "..") {
         if (!is_dir($dirname."/".$file))
            unlink($dirname."/".$file);
         else
            delete_directory($dirname.'/'.$file);    
      }
   }
   closedir($dir_handle);

    $htmlFormat = new RenderAsHTML();

  $htmlFormat->StreamRoot = './images/';
 
  $result_html = $rs->Render2($htmlFormat,PageCountModeEnum::$Estimate,$Extension,$MimeType,$Encoding,$Warnings,$StreamIds);

      $i = 0;

      // handle the images by writing them to the temp images folder
      foreach ($StreamIds as $StreamId){

$result_image = $rs->RenderStream($htmlFormat,$StreamIds[$i],$Encoding,$MimeType);

if (!$handle = fopen("./images/" . $StreamIds[$i] , 'wb'))
{
  echo "Cannot open file for writing output";
exit;
}
if (fwrite($handle, $result_image) === FALSE)
{
echo "Cannot write to file";
exit;
}
fclose($handle);
$i++;
}
// construct the html of the report
$content="<div align='center'>";

$content .="<style type='text/css'> td{border-top-width: 1px;border-top-style: solid;border-top-color: transparent;}</style>";
$content .="<div>";

$content .='<div>';
/* add the report content */
$content .=$result_html;
$content .='</div>';
$content .='</div>';

$content .='</div>';

Note that the report is rendered as HTML by the Render2 method of the SSRSReport class provided by the SQL Server Reporting Services SDK for PHP . If the report contains any images or charts, the method generates image files for these and stores them in a specified folder – hence the requirements to clear out the image folder before rendering the report to remove any images left over from previously viewed reports. One gotcha that we found when implementing the solution was that these images are generated with no file extension, so we had to configure IIS to map the .* file name pattern to the Image/png MIME type.

The solution we built was based on pre-beta releases of Drupal 7 and the related SQL Server driver for PDO and SQL Server module. Given the complex dependencies on multiple pre-release software components, it seems foolhardy to publish the full source code for this solution until all of the dependencies are stabilized in a final release.  At that point, the plan is to update the code for our solution and publish it on CodePlex so that anyone can examine and extend it, so watch this space!

Thursday 18 November 2010

First Look at SQL Server Denali (Part 2 – Snippets)

This is a quick post for a new feature I discovered while writing the first “Denali” post. SQL Server “Denali” CTP1 is available for download from the Microsoft Download Center here.

The first post used a very straightforward While loop to demonstrate the improved debugging ability. You now have the ability to add code Snippets and Surround With Snippets.

If I use the code from the previous post without the loop it looks like this:

image

If you select and right-click the UPDATE statement you are presented with a dialog box which includes Insert Snippet and Surround With. By selecting Surround With you are presented with a further menu which comprises Begin, If, and While. You can also type your own freeform text. For this example we need a while loop and so you click While and press Enter. The statement is now surrounded by a WHILE snippet as shown:

image

It is now a short step to complete the statement and finish the code as follows:

image

It could be argued that this doesn’t save a lot of time because it has only typed three words, however it does prevent mistakes and correctly formats your SQL.

The snippets feature is similar except you are typically using it for new queries, or new statements in a query. In the previous example you might want to check the results of the query and because you run this often you have decided to create a view (it is at this point that I wish I created a query which had a purpose rather than something which just looped around for the sake of it. Anyway, I digress…).

By creating a new query, right-clicking and selecting Insert Snippet, you are presented with these options:

image

Scrolling down reveals View and when you click this the following snippet is created:

image

From here, you can quickly finish the syntax to arrive at the final statement:

image

Again, nothing amazing, but there is a wide range of snippets and I wouldn’t be surprised if this was expanded for release. This is actually a feature that comes from Visual Studio and therefore we also have the Code Snippets Manager which I personally think makes these features much more useful. If you go to the Tools menu and select Code Snippets Manager, you are presented with the following dialog box which allows you to add your own snippets:

image

Adding a snippet is not as straightforward as clicking Add and typing some SQL however. Snippets are XML files with a .snippet extension. If you want to create your own I would recommend searching for .snippet files and using an existing one as a starting point. The Create View snippet has the following syntax:

image

A bit of editing (and bad highlighting) gets you here:

image

If you now save this in the Snippets folder in My Documents with a .snippet extension, you are ready to add it. Return to Code Snippet Manager and click Add. Select your snippets folder and it will appear as a folder in the snippet list. Now you can add a snippet in the standard way, but have your own snippet:

image

Now there are real benefits from snippets by creating your own standardized SQL. I’m not aware of a more user-friendly method to create the snippets, but if anyone else has found one, I’d love to hear about it in the comments.

Wednesday 17 November 2010

First Look at SQL Server Denali (Part 1 – Debugging)

This is a first look at some of the new features in SQL Server “Denali” (which will most likely become SQL Server 2011).

SQL Server “Denali” CTP1 is available for download from the Microsoft Download Center here.  The press release is here.

image

Today, I’ll be focussing on the improved debugging functionality in SQL Server Management Studio. Following is a straightforward query to loop through the SalesOrderDetail table 10 times and add one to the UnitPrice each time (I’m not too concerned in this example with why you would want to do that!). I have filtered the query to rows which have a SalesOrderDetailID of less than 10.

image

We can now add a breakpoint on the UPDATE statement.

image

So far, nothing has changed, but if you right-click the breakpoint you can add limit the breakpoint by Location, Condition, and Hit Count, specify a filter for certain processes and specify what to do when a breakpoint is hit.

image

In this example, I have added a Condition of 2’">2’">2’">‘@counter>2’ and a HitCount of 2. Therefore the break should occur the second time the @counter value is over 2. If we run the code in debug mode we can see (after moving the panes around to get it in one screen capture) that value of @counter is 4 which is indeed the second time it has hit a value over 2.

image

This is a very straightforward example, but you can see that this makes the ability to use breakpoints much more powerful and therefore much more useful.

Tuesday 16 November 2010

The Async Modifier and the Await Operator - New Features in the Next Edition of C#

At the recent Professional Developers' Conference, Anders Hejlsberg gave a presentation that demonstrated some of the new language features that Microsoft are working on as part of the next version of Visual C# and Visual Basic. The thrust of the next build of these languages is oriented towards making it easier to write asynchronous code. These new features add native functionality to Visual C# and Visual Basic that are built on top of the TPL, introduced in the .NET Framework 4.0. This article concentrates on the syntax extensions proposed for C#, although similar extensions will be added to Visual Basic.

The purpose of the TPL is to make it easier to build applications that can take advantage of the now-commonplace crop of multicore processors to implement concurrency and multithreading. To recap, using the TPL you can implement a concurrent operation by defining a Task object that references the code to be run concurrently. When you execute the task by using the Start method, the .NET Framework uses its own scheduling algorithm to allocate the Task to a thread and set this thread running at a time convenient to the operating system, when sufficient resources are available. This level of abstraction frees your code from the requirement to understand and manage the workload of your computer (for more information about Tasks, see the post Make Your Applications More Efficient with MultiTasking.)

If you need to perform another operation when a specific task completes, you have a couple of choices:

  1. You can manually wait for the task to complete by using one of the Wait methods exposed by the Task type. You can then initiate the new operation, possibly by defining another task.

  2. You can define a continuation. A continuation simply specifies an operation to be performed when a given task completes. The .NET Framework automatically executes the continuation operation as a task that it schedules when the original task finishes. For information about continuations, see the MSDN article TPL and Traditional .NET Asynchronous Programming.

However, although the TPL provides the Task type as an abstraction for a concurrent operation, it is still often necessary to write potentially awkward code to solve some of the common problems that developers frequently encounter when building applications that contain sections that may run concurrently. Additionally, bear in mind that the purpose of the TPL is to implement concurrency (typically by running multiple threads) whereas asynchronicity is a subtley different problem that may require coordinating operations that run on a single thread. This is where the async modifier and the await operator come into play. As an example, consider the following simple WPF window (the XAML definiton is also shown).



<Window x:Class="TestApp.MainWindow"
xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
Title="Asynchronous Demonstration" Height="204" Width="525" ResizeMode="NoResize">
<Grid>
<Button Content="Perform Work Synchronously" Height="23" HorizontalAlignment="Left" Margin="15,32,0,0" Name="synchronous" VerticalAlignment="Top" Width="187" Click="synchronous_Click" />
<Rectangle Height="142" HorizontalAlignment="Left" Margin="214,12,0,0" Name="rectangle1" Stroke="Black" VerticalAlignment="Top" Width="280" />
<Label Content="Status:" Height="28" HorizontalAlignment="Left" Margin="223,27,0,0" Name="label1" VerticalAlignment="Top" />
<TextBox Height="23" HorizontalAlignment="Left" Margin="286,27,0,0" Name="status" VerticalAlignment="Top" Width="195" IsEnabled="False" />
<Label Content="Duration:" Height="28" HorizontalAlignment="Left" Margin="223,61,0,0" Name="label2" VerticalAlignment="Top" />
<TextBox Height="23" HorizontalAlignment="Left" IsEnabled="False" Margin="286,61,0,0" Name="duration" VerticalAlignment="Top" Width="195" />
<Button Content="Perform Work using Async Task" Height="23" HorizontalAlignment="Left" Margin="15,74,0,0" Name="taskBasedAsync" VerticalAlignment="Top" Width="187" Click="taskBasedAsync_Click" />
<Button Content="Perform Work using C# Async" Height="23" HorizontalAlignment="Left" Margin="15,114,0,0" Name="nativeAsync" VerticalAlignment="Top" Width="187" Click="nativeAsync_Click" />
</Grid>
</Window>

This window defines three buttons that each perform the same long-running operation, but by using a different mechanism. The first button performs the work synchronously, in a naive manner, as shown in the following code:

public partial class MainWindow : Window
{
...
private void synchronous_Click(object sender, RoutedEventArgs e)
{
duration.Clear();
status.Text = "Running - Synchronous";
Stopwatch watch = Stopwatch.StartNew();
DoWork();
duration.Text = string.Format("{0} ms", watch.ElapsedMilliseconds);
status.Text = "Completed - Synchronous";
}
...
private void DoWork()
{
Thread.SpinWait(Int32.MaxValue / 2);
}
...
}

This code displays the text "Running – Asynchronous" in the status text box in the right-hand portion of the window before starting a System.Diagnostics.Stopwatch object (the code uses this object to time the duration of the operation). The code then calls the DoWork method, which simply performs a busy wait for a few seconds to simulate a long-running operation, before displaying the time taken and updating the text in the status text box to "Completed – Synchronous". When you run this code, if you click the "Perform Work Synchronously" button, you should observe the following phenomena:

  • The Window becomes unresponsive while the DoWork method runs; you cannot even move the window.

  • The status text box never actually displays the message "Running – Synchronous". The contents of this text box are only updated when the synchronous_Click method completes, after the contents of the status text box has been overwritten with the message "Completed - Synchronous".

Both of these issues are related to the same problem; the thread running the event loop that waits for Windows messages and handles them is kept busy by the DoWork method, and cannot process any events (such as requests to move the Window) until the synchronous_Click method finishes. Similarly, Windows cannot update the display until the thread running the event loop has finished the synchronous_Click method, so the status text box never displays the text "Running – Synchronous", only the message "Completed - Synchronous".

The usual strategy to avoid these problems and ensure that the user interface remains responsive while the application performs a long-running operation is to run this operation on a separate thread, thus freeing the thread running the Windows event loop to process other messages. You can achieve this with the TPL by using code such as the taskBasedAsync_Click method in the following example. This is the code for the "Perform Work using Async Task" button shown in the earlier image:

public partial class MainWindow : Window
{
...
private void taskBasedAsync_Click(object sender, RoutedEventArgs e)
{
duration.Clear();
status.Text = "Running - Async, Task-Based";
Task t = new Task(() => DoWork());
Stopwatch watch = Stopwatch.StartNew();
t.ContinueWith((task) => UpdateStatus(watch, "Completed - Async, Task-Based"));
t.Start();
}
...
private void DoWork()
{
Thread.SpinWait(Int32.MaxValue / 2);
}

private void UpdateStatus(Stopwatch watch, string message)
{
this.Dispatcher.Invoke(new Action(() =>
{
duration.Text = string.Format("{0} ms", watch.ElapsedMilliseconds);
status.Text = message;
}), DispatcherPriority.ApplicationIdle);
}
...
}

When you run the application and click the "Perform Work using Async Task" button, the user interface remains responsive, even while the DoWork method runs. This is because the DoWork method is performed by a task on a separate thread. Additionally, the status text box successfully displays the message "Running - Async, Task-Based" while the DoWork method runs. The tricky part is arranging for the display to be updated with the time taken to complete the DoWork method and the message " Completed - Async, Task-Based". This is achieved with a continuation that runs when the task performing the DoWork method completes. However, because this task runs on a different thread from that responsible for managing the user interface, it cannot directly modify the properties of the duration and status text boxes, but has to queue updates to the WPF Dispatcher object for the Window instead. When the thread handling the Windows event loop is free, it runs the code specified by the Dispatcher.Invoke method.

So, although this code works, it is messy; if you want to update the user interface, you have to define a continuation that invokes the Dispatch.Invoke method because the continuation does not run on the thread that owns the user interface elements. This is where the async modifier and await operator can prove very useful.

The async modifier to a method indicates that the method contains functionality that can be run asynchronously. The await operator specifies the point at which asynchronous operations can begin inside an async method. As a final example, consider the nativeAsync_Click method shown below. This method runs when the user clicks the "Perform Work using C# Async" button.

public partial class MainWindow : Window
{
...
private async void nativeAsync_Click(object sender, RoutedEventArgs e)
{
duration.Clear();
status.Text = "Running - Async, Native C#";
Task t = new Task(() => DoWork());
Stopwatch watch = Stopwatch.StartNew();
t.Start();
await t;
duration.Text = string.Format("{0} ms", watch.ElapsedMilliseconds);
status.Text = "Completed - Async, Native C#";
}
...
private void DoWork()
{
Thread.SpinWait(Int32.MaxValue / 2);
}
...
}

This code looks like a hybrid combination of the first two examples. It displays the message "Running – Async, Native C#" in the status text box before creating and running a task to perform the DoWork method on a separate thread. However, the await operator is where the clever stuff kicks in. Notice that the nativeAsync_Click method is defined with the async modifier. At run-time, the await operator causes the async method to return immediately from whence it was called without waiting for any subsequent code to run. When the subject of the await operator (the task running the DoWork method) completes, the nativeAsync_Click method resumes running at this point and the statements following the await operator are performed, updating the display with the time taken to perform the operation and the message " Completed - Async, Native C#". In fact, this magic is nothing more than an exercise in the reworking of your code by the C# compiler. When the C# compiler encounters the await operator in an async method, it effectively reformats the code that follows this operator as a continuation that runs on the same thread as the async method. And because the thread that was running the async method was the thread running the Windows event loop, it has direct access to the controls in the Window and can update them directly without routing them through the WPF Dispatcher object for the Window.

Although this approach looks quite simple at first glance, it is important to bear in mind a few points to avoid some possible misconceptions:

  • The async modifier does not signify that a method runs asynchronously on a separate thread. All it does is specify that the code in the method can be divided into one or more continuations. When these continuations run, they execute on the same thread as the original method call.

  • The await operator specifies the point at which the C# compiler can split the code into a continuation. The await operator itself expects its operand to be an awaitable object. An awaitable object is a type that provides the GetAwaiter method which returns an object that in turn provides the BeginAwait and EndAwait methods. The C# compiler converts your code into statements that uses these methods to create an appropriate continuation. The Visual Studio Async CTP provides extension methods for the Task class, and the code shown in the example above invokes the await operator on a Task object.

To conclude, the async modifier and await operator are powerful constructs that enable you to simplify asynchronous code, and as such they are natural partners with the TPL. Using the async and await keywords, you can concentrate on the logic of your application and let the compiler worry about how to divide your code into one or more continuations.

Thursday 11 November 2010

What are Microsoft’s real virtualization technologies?

OK, so something is either real or virtual, right? But behind the contradiction implied by this title, is a serious question. What exactly is virtualization?

Not only are virtualization technologies becoming more and more prevalent in all kinds of organizations, from small businesses through to global enterprises, but the term "virtualization" is also being used more and more, and sometimes to describe things that may not be virtualization as this term has commonly been understood. For example Microsoft, and other vendors, has applied the "virtualization" brush to a range of technologies, not all of which are really virtual! Or have they …

One definition of virtualization is: The capability to run an application or a computer in a virtual environment without affecting the components that already exist on that particular desktop or server.

This definition mostly works for desktop virtualization technologies, such as Virtual PC 2007 and Windows Virtual PC on Windows 7, VMware Workstation, and server virtualization technologies such as Microsoft Hyper-V and VMware Server, that enable several operating systems to co-exist on the same physical hardware, and in protected "bubbles" so that one OS does not affect any others.

However, as we shall see when we go through some of Microsoft's current "virtualization" technologies, the above definition doesn't always fully explain things.

Server virtualization is probably the least ambiguous technology area for Microsoft. Server virtualization enables multiple servers, such as domain controllers, DNS servers, and Web servers, to run as discrete virtual machines on the same physical computer host. Server virtualization enables the physical consolidation of multiple traditional servers to improve and maximize hardware utilization, and to consolidate workloads onto fewer enterprise class servers. Server virtualization is also about reducing costs, by using fewer physical computes, with reduced electrical costs for cooling and server power consumption.

But what about desktop virtualization? Desktop virtualization enables software applications, operating system, and hardware configuration to be packaged to run in a self-contained, virtual environment, enabling users to run multiple operating systems and applications on a single computer. With desktop virtualization technologies, such as Microsoft Virtual PC 2007 SP1 and Windows Virtual PC, including Windows 7's XP Mode, the virtual machine files are installed and executed on the local client desktop computer. In enterprise environments, desktop virtualization can be augmented with technologies such as MED-V that are used to deliver and manage client virtual machines over the network from central servers.

Isn't this clear and unambiguous? Yes, if you just include "traditional" definitions of desktop virtualization, but what about approaches such as using Microsoft's Remote Desktop Services (RDS) to deliver desktops and applications hosted on central servers, but presented on user's desktops as if they were running locally? This approach de-couples the user desktop from their physical workstation, but importantly there are not separate virtual machines on the server; rather every user is running a session on the server, and any applications they run must be compatible with the server OS. This approach is sometimes called presentation or session virtualization. And where shortcuts to remote applications are provided on user's desktops, this approach is sometimes also referred to as a type of application virtualization.

And then there is the whole area of Virtual Desktop Infrastructure (VDI) …. VDI is an alternative server-based desktop virtualization method that uses virtualization technologies enable users to run their own client desktops, but hosted on a central server. In Microsoft's VDI, multiple client operating systems run in virtual machines on a Hyper-V server; the server then remotely presents each desktop to the client computer, over RDS. By using virtual machines running on the server, every user can run their own desktop OS version (and not be limited to a server OS), and customize their own unique desktop settings.

So, there's more to virtualization than virtual machines; we've also got session-based virtualization and, perhaps most importantly, there's also the concept of de-coupling, where elements such as hardware, OS, applications, and user state and data, that in a traditional rich desktop environment would be all packaged together, are now potentially in their own separate spaces.

Does considering virtualization to be of different types help with some of the other "virtualization" technologies?

Let's consider user state virtualization. The term user state virtualization describes the ability to separate user data and settings from applications and operating systems, so that users can access their own personalized environment, no matter which device they are currently using. So, in this case, it is the de-coupling element of virtualization that is key – we are not talking about multiple concurrent user states here, and there are no virtual machines involved. Instead there's the ability to separate user state data from the applications and operating system so that users can move from environment to environment, or device to device, and automatically pick up their saved data, favorites, application settings and so on no matter where they are. In Windows environments, user state virtualization technologies include Folder Redirection, Offline Files, and Roaming User Profiles.

How about application virtualization? As a generic term application virtualization describes technologies that deliver applications across the network to end-users, without the need to install these applications on the user's desktop. There are two key methods for virtualizing applications; through specific application virtualization tools and through presentation virtualization.

Application virtualization tools typically package a software application so that it runs in a self-contained, virtual environment. This virtual environment contains all the information required to run the application on the client computer, without installing the software application locally. So, again the decoupling element of virtualization is important, as the application is not "installed" on the host operating system. In typical implementations, such as Microsoft's App-V, application code is downloaded on demand from a central server and executed on the local desktop computer. Although the application is running within a virtual OS environment, this is not a complete virtual machine; rather it provides virtualized interfaces to OS components such as the Windows Registry, and is a great approach for dealing with incompatibilities between applications; but is not useful for dealing with incompatibilities between an application and the host OS; this is because the application it does still make use of OS features, and must still meet the same platform and system requirements as a regular application; for example, if an application won't run under Windows 7 as a traditional installed application, it will still fail when run under App-V on a Windows 7 workstation.

Presentation virtualization typically describes technologies where the execution of application code takes place remotely from the end-user's interaction with the application interface. In Windows environments, presentation virtualization is used by applications delivered using RDS, such as traditional terminal services applications. Using technologies such as Microsoft's RemoteFX, such applications can appear to the end-user as if they are running locally, particularly if configured as a RemoteApp, where the individual application is launched from a regular application shortcut on the user desktop; in this way the user may be unaware that the application is actually a terminal into a remote session running on a server. So, again we've got de-coupling, and with sessions rather than virtual machines.

Looking at Microsoft as an example vendor, what we need to do is think of different types or classes of virtualization. The key feature in all types of virtualization is perhaps the de-coupling of elements such as hardware, OS, applications, and user state and data. Using this as a principle, we can see that the "traditional" approach, such as desktop and server virtualization, achieves this de-coupling by using distinct virtual machines. With user state virtualization, the de-coupling is achieved using remote central data storage; with application virtualization, the de-coupling can be from virtual application environments or through sessions in the case of remote applications and desktops. And implementations using VDI, can achieve de-coupling by using all these tools; virtual machines, centralized storage, and session-based technologies.

So, as long as we can think of the core virtualization concept as a separating or de-coupling approach, then there's no reason not to use the term for a wide range of approaches – and if this is case maybe all Microsoft's virtualizations are real ….