Nevron was kind enough to show me their chart and gauge for Reporting Services. In certain areas, their products exceed the charting capabilities of Reporting Services. Here are the highlights:
The product provides support for SQL Server Reporting Services (SSRS) 2005, 2008, 2008R2 and SQL Server Data Tools (SSDT) 2012.
Advanced Chart Designer
Complete set of 2D and 3D Charting Types
Support for Code Customization – Nevron Chart for Reporting Services supports customization through C# code, which allows you to use the full Nevron Chart for .NET API in SSRS and achieve a higher degree of customization.
Expressions Everywhere – Nearly all properties of the Chart report item can now be specified by expressions. Appearance styles (fill, stroke, shadow and text) are also reworked to support both constant specification and specification that is based on expressions. This allows you to bind nearly all configurable aspects of the report item to user defined parameters and/or data.
Support for 2D/3D Combo Charts – Combo Charts are used to combine certain series types within the same chart area.
XML Formatted Texts – This feature is applicable to all elements which display text. It allows you to mix fonts, mix bold, italic underline formatting and many other advanced features like different fill styles, shadows, image filters etc. – all inside a single label, title etc.
Non-overlapping Data Labels Layout – All Cartesian Chart types support automatic non-overlapping data labels layout. The data label layout prevents labels from overlapping in both ordinal and scatter charts and works in 2D and 3D appearance modes.
Advanced Axis Model – complete set of axis-related features designed to target even the most compelling charting requirements.
Suppose your company has a web application that allows customers to enter comments, such as to provide feedback about their product experience. How do you derive knowledge from these comments? Perhaps, customers are complaining a lot about a particular product. Or, they are raving about your service. Enter semantic search – a new feature in SQL Server 2012.
Building upon full text search, semantic search allows you to search not only for words, but also for the meaning (semantics) of these works. Mark Tabladillo, Ph.D., gave us great presentation at our Atlanta BI January meeting. He demoed how semantic search can be used to find the most common phrases in a document and how to find similar documents from a given search criteria.
Besides external documents, you can apply semantic search to text data already stored in a database, such as a text-based Comments field. Again, this is possible because semantic search builds upon the full-text search capabilities of SQL Server. So, any column that supports full-text search can be enhanced with semantic search.
Going back to my scenario, here is what it takes to configure a text field for semantic search:
When you use semantic search, SQL Server performs statistical analysis of the words in the column contents. This analysis requires base data that is provided as a SQL Server database. Because the database is not installed by default, as a perquisite of using semantic search, you need to run the semantic database installer. The installer can be found on the SQL Server setup disk in the following folders:
For x86, the setup is \x86\Setup\SemanticLanguageDatabase.msi
For x64, the setup is \x64\Setup\SemanticLanguageDatabase.msi
The setup simply extracts the MDF and LDF files of the semantic database to a folder that you specify during the installation. Next, simply attach the semantics database to your SQL Server 2012 instance as you would with any other database.
The next step is another step that you need to perform only once for each SQL Server instance. Register the semantics database using this command:
This command creates a semantic-enabled full-text index on the Comments column in the Production.ProductReview table on the AW2008FullTextCatalog full text catalog. The AW2008FullTextCatalog text catalog is included with the AdventureWorks2012 database. The CHANGE_TRACKING = AUTO clause instructs SQL Server to auto-update the index, and STOPLIST = SYSTEM specifies that the default full-text system STOPLIST should be used to parse words.
5. Now that you’ve done the ground work, you can perform the semantic search. For example, the following query returns the most frequent words for each document:
Note that the document_key matches the ProductReviewID column (primary key) of the [Production].[ProductReview] table allowing you to match the semantic search results to the document. For example, “socks” and “trip” were found in the row with the primary key of 1 whose Comments column contains the following text:
“I can’t believe I’m singing the praises of a pair of socks, but I just came back from a grueling 3-day ride and these socks really helped make the trip a blast. They’re lightweight yet really cushioned my feet all day. The reinforced toe is nearly bullet-proof and I didn’t experience any problems with rubbing or blisters like I have with other brands. I know it sounds silly, but it’s always the little stuff (like comfortable feet) that makes or breaks a long trip. I won’t go on another trip without them!”
You can feed these results to a data mining model if you want to learn which phrases customers tend to use together similar to identifying what products customers tend to purchase together (market basket analysis). “Together” here means within a specific document.
What if you want to get the most popular phrases across all comments? I think the following query should help although there might be a better way. The query groups phrases and sorts them in a descending order by the average score. SELECTTOP(5)KEYP_TBL.keyphrase,AVG(score)AS Score
MVP fellow, Andrew Burst, discusses Gartner’s newly released DW (not BI, mind you) magic quadrant. It looks like Gartner fell in love with Teradata but Microsoft is moving up as well thanks to the advances in its Parallel Data Warehouse (PDW) and in-memory offerings.
I ran into a severe performance issue with the Aggregate function and unrelated dimensions which I reported on connect. In this context, an unrelated dimension is a dimension that doesn’t join any measure group in the cube. Why have unrelated dimensions? My preferred way to implement time calculations is use a regular dimension that is added to the cube but not related to any measure groups. Then, I use scope assignments to overwrite the time calculations, e.g. for YTD, MTD, etc. For example, the scope assignment for YTD might look like:
Notice the use of the Aggregate function which when executed maps to the default aggregation function of the underlying measure. For some reason with SQL Server 2012, a query that uses the Relative Date dimension experiences a significant performance hit. Replacing Aggregate with Sum fixes the issue, assuming you can sum up the affected measure to produce the time calculations.
https://prologika.com/wp-content/uploads/2016/01/logo.png00Prologika - Teo Lachevhttps://prologika.com/wp-content/uploads/2016/01/logo.pngPrologika - Teo Lachev2013-02-05 13:55:462016-02-16 04:57:16Performance Degradation with the Aggregate Function and Unrelated Dimensions
What options does a PowerPivot user have to refresh data in a PowerPivot model on the desktop, aka PowerPivot for Excel?
Prior to Excel 2013, the answer was just one – manual refresh by either clicking the Refresh button in the PowerPivot window or the Refresh button in the Existing Connections dialog box. Note that the Refresh button in the Excel ribbon doesn’t work since Excel doesn’t know anything about PowerPivot. Not does the checking the “Refresh data when opening the file” checkbox in the PowerPivot connection. In Excel 2010, these options won’t reopen the PowerPivot connections to the data sources. Instead, the net effect is that they will simply refresh the pivot reports from the pivot cache which is not what you’re after.
Starting with Excel 2013, however, Excel and PowerPivot play better together as I discussed in my What’s New in Office 2013 BI blog. And, now we have additional options to refresh data:
Right-click the pivot report, go to PivotTable Options, click the Data tab, and then click “Refresh data when opening the file”. Enabling this option will refresh the data in the PowerPivot tables that are used on the report when you re-open the Excel workbook file.
In the Excel Data ribbon, click Connections, select the connection you want to refresh, and then click Properties. Notice that all PowerPivot connections are now exposed in the Connections dialog. In the Connections Properties dialog box, check the “Refresh data when opening the file”. This option will open the connection when you open the file, and refresh all PowerPivot tables using the connection.
3. Finally, you can also automate PowerPivot tasks, including refreshing data. For example, if you want to refresh the ResellerSales table on open, you can add the following line to the Workbook_Open() event:
Speaking of automation, everything you can do in the Excel UI is exposed in the object model and automatable in Excel 2013. Here is another example of adding a table from the Excel workbook to the model and then creating a relationship between that table and another table which is already in the model:
Besides the usual hassle configuring PerformancePoint, including insufficient permissions to databases and service accounts, version 2013 requires the SQL Server 2008 R2 drivers. This is surprising considering that SharePoint 2013 shipped after SQL Server 2012.
In a previous blog, I explained how to configure the SharePoint 2013 BI Center. When you go to any of the PerformancePoint-related links, such as Dashboards or PerformancePoint Content, you’ll see the following ribbon.
The Dashboard Designer button is the new way to launch the PerformancePoint Designer. However, when you attempt to configure a data source pointing to Analysis Services, you will likely get an error. The first stop for troubleshooting SharePoint and PerformancePoint issues is of course the Windows Event Log. If you examine the Windows Event Log, you’ll see that PerformancePoint fails to load the 10.0 version of Microsoft.AnalysisServices.AdomdClient. This is the version that’s included in SQL Server 2008 R2. This sends you to the SQL Server 2008 R2 Feature Pack page, from where you can download and install the Microsoft SQL Server 2008 R2 ADOMD.NET library. Now, you can connect to Analysis Services.
The next trip to the SQL Server 2008 R2 Feature Pack page will happen when you try to import (not create) KPIs defined in an Analysis Services cube. This time the error in the Event Log indicates that PerformancePoint requires the 10.0 version of the Microsoft.AnalysisServices dll, which represents the Analysis Services Management Objects (AMO). Back to the SQL Server 2008 R2 Feature Pack, you need to download and install Microsoft SQL Server 2008 R2 Analysis Management Objects. While you there, you might as well download and install Microsoft Analysis Services OLE DB Provider for Microsoft SQL Server 2008 R2 although I don’t think PerformancePoint uses it.
Where should I put my BI reports? Should I upload them to department-level SharePoint sites or put them in one place?
These are common questions that we get from customers. Delivering on the promise of pervasive BI, my preference is to centralize BI artifacts in a single place. Ideally, this BI depository should be the SharePoint Business Intelligence Center. If organizational security is required, you can control security at SharePoint site or library level. For example, you can create department-specific PowerPivot galleries.
The BI Center is one of the SharePoint site templates that is specifically designed to host BI reports. In SharePoint 2013, Microsoft has extended the BI Center to accommodate various types of BI documents.
When I first installed the BI Center in SharePoint 2013 (click Settings (the wheel in the top-right corner), Site Contents, New Subsite, Enterprise Tab, Business Intelligence Center), I was confused. The images are not clickable and the default home page doesn’t offer much more. It turned out that by default, the BI Center doesn’t add a navigation menu. To fix this:
Navigate to the BI Center, navigate to the BI center link.
Click Settings, Site Settings, and click the Navigation link (under the Look and Feel section).
In the Current Navigation section, select the Structural Navigation option.
In the Structural Navigation Sorting section, click Add Heading to add a new menu item for each library your users want to navigate to. To get the links, back to the BI Center, go to Settings, Site Contents, and then right-click the library you are interested in, such as Dashboards, and click Copy Shortcut. Then, paste the shortcut in the URL field in the Navigation Heading dialog box.
This is what the resulting navigation pane might look like. One thing that might not be obvious is that the added benefit of creating your links using the SharePoint structural navigation is that links reflect security. For example, if the user doesn’t have permission to a library, the user won’t see the link. Note that there might be additional steps required, such as to enable the library content types, as with SharePoint 2010.
https://prologika.com/wp-content/uploads/2016/01/logo.png00Prologika - Teo Lachevhttps://prologika.com/wp-content/uploads/2016/01/logo.pngPrologika - Teo Lachev2013-02-04 01:23:002016-02-16 05:57:42The SharePoint 2013 Business Intelligence Center
In a previous blog “Installing HDInsight Server for Windows”, I introduced you to the Microsoft HDInsight Server for Windows. Recall that HDInsight Server for Windows is a Windows-based Hadoop distribution that offers two main benefits for Big Data customers:
An officially supported Hadoop distribution on Windows server – Previously, you can set up Hadoop on Windows as an unsupported installation (via Cygwin) for development purposes. What this means for you is that you can now set up a Hadoop cluster on servers running Windows Server OS.
Extends the reach of the Hadoop ecosystem to .NET developers by allowing them to write MapReduce jobs in .NET code, such as C#.
And, in previous blogs, I’ve introduced you to Hadoop. Recall that there are two main reasons for using Hadoop for storing and processing Big Data:
Storage – You can store massive files in a distributed and fault-tolerant file system (HDFS) without worrying that hardware failure will result in a loss of data.
Distributed processing – When you outgrows the limitations of a single server, you can distribute job processing across the nodes in a Hadoop cluster. This allows you to perform crude data analysis directly on files stored in HDFS or execute any other type of jobs that can benefit from a parallel execution.
This blog continues the HDInsight Server for Windows journey. As many of you probably don’t have experience in Unix or Java, I’ll show you how HDInsight makes it easy to write MapReduce jobs on a Windows machine.
Note Writing MapReduce jobs can be complex. If all you need is performing some crude data analysis, you should consider an abstraction layer, such as Hive, which is capable for deriving the schema and generating the MapReduce jobs for you. This doesn’t mean that experience in MapReduce is not useful. When processing the files go beyond just imposing a schema on the data and querying the results , you might need programming logic, such as in The New York Times Archive case.
As a prerequisite, I installed HDInsight on my Windows 8 laptop. Because of its prerelease status, the CTP of HDInsight Server for Windows currently supports a single node only which is fine for development and testing. My task is to analyze the same dataset that I used in the MS BI Guy Does Hadoop (Part 2 – Taking Hadoop for a Spin) blog. The dataset (temp.txt) contains temperature readings from weather stations around the world and it represents the weather datasets kept by National Climatic Data Center (NCDC). You will find the sample dataset in the source code attached to this blog. It has the following content (the most important parts are highlighted in red: the year found in offset 15 and temperature found in offset 88).
Note that the data is stored in its raw format and no schema was imposed on the data. The schema will be derived at runtime by parsing the file content.
Installing Microsoft .NET SDK for Hadoop
The Microsoft .NET SDK for Hadoop facilitates the programming effort required to code MapReduce jobs in .NET. To install it:
Install NuGet first. NuGet is a Visual Studio extension that makes it easy to add, remove, and update libraries and tools in Visual Studio projects that use the .NET Framework.
Open Visual Studio (2010 or 2012) and create a new C# Class Library project.
Go to Tools ð Library Package Manager ð Package Manager Console.
In the Package Manager Console window that opens in the bottom of the screen, enter: install-package Microsoft.Hadoop.MapReduce –pre
This command will download the required Hadoop binaries and add them as references in your project.
Coding the Map Job
The Map job is responsible for parsing the input (the weather dataset), deriving the schema from it, and generating a key-value pair for the data that we’re interested in. In our case, the key will be the year and the value will be the temperature measure for that year. The Map class derives from the MapperBase class defined in Microsoft.Hadoop.MapReduce.dll.
At runtime, HDInsight will parse the file content and invoke the Map method once for each line in the file. In our case, the Map job is simple. We parse the input and extract the temperature and year. If the parsing operation is successful, we return the key-value pair. The end result will look like this:
(1950, 0)
(1950, 22)
(1950, −11)
(1949, 111)
(1949, 78)
Coding the Reduce Job
Suppose that we want to get the maximum temperature for each year. Because each weather station might have multiple readings (lines in the input file) for the same year, we need to combine the results and find the maximum year. This is analogous to GROUP BY in SQL. The following Reduce job gets the work done:
The Reduce job is even simpler. The Hadoop framework pre-processed the output of the Map jobs before it’s sent to the Reduce function. This processing sorts and groups the key-value pairs by key, so the input to the Reduce job will look like this:
(1949, [111, 78])
(1950, [0, 22, −11])
In our case, the only thing left for the Reduce job is to loop through the values for a given key (year) and return the maximum value, so the final output will be:
(1949, 111)
(1950, 22)
Testing MapReduce
Instead of deploying to Hadoop each time you make a change during the development and testing lifecycle, you can add another project, such as a Console Application, and use it as a test harness to test the MapReduce code. For your convenience, Microsoft provides a StreamingUnit class in Microsoft.Hadoop.MapReduce.dll. Here is what our test harness code looks like:
The code uses a test input file. It reads the content of the file one line at the time and adds each line as a new element to an instance of ArrayList. Then, the code calls the StreamInsight.Execute method to initiate the MapReduce job.
Deploying to Hadoop
Once the code is tested, it’s time to deploy the dataset and MapReduce jobs to Hadoop.
Deploy the file to the Hadoop HDFS file system. C:\Hadoop\hadoop-1.1.0-SNAPSHOT\bin>hadoop fs -copyFromLocal D:\MyApp\Hadoop\MapReduce\temp.txt input/Temp/input.txt
Note When you execute the hadoop command shell in the previous step, the file will be uploaded to your folder. However, if you use the JavaScript interactive console found in the HDInsight Dashboard, the file will be uploaded to the Hadoop folder in HDFS because the console runs under the hadoop user. Consequently, the MapReduce job won’t be able to find the file. So, you use the hadoop command prompt.
2. Browse the file system using the web interface (http://localhost:50070) to see that the file is in your folder.
3. Finally, we need to execute the job with HadoopJobExecutor, which be called in various ways. The easiest way is to use MRRunner D:\MyApp\Hadoop\MapReduce\FirstJob\bin\Debug>.\mrlib\mrrunner -dll FirstJob.dll
Job\bin\Debug\MRLib\Microsoft.Hadoop.CombineDriver.exe, D:\MyApp\Hadoop\MapReduce\FirstJob\bin\Debug\FirstJob.dll, D:\MyApp\Hadoop\MapReduce\FirstJob\bin\Debug\Microsoft.Hadoop.MapReduce.dll, D
12/12/28 12:35:20 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform… using builtin-java classes where applicable
log4j:ERROR Failed to rename [C:\Hadoop\hadoop-1.1.0-SNAPSHOT\logs/hadoop.log] to [C:\Hadoop\hadoop-1.1.0-SNAPSHOT\logs/hadoop.log.2012-12-27].
12/12/28 12:35:20 WARN snappy.LoadSnappy: Snappy native library not loaded
12/12/28 12:35:20 INFO mapred.FileInputFormat: Total input paths to process : 1
12/12/28 12:35:20 INFO streaming.StreamJob: getLocalDirs(): [c:\hadoop\hdfs\mapred\local]
12/12/28 12:35:20 INFO streaming.StreamJob: Running job: job_201212271510_0010
12/12/28 12:35:20 INFO streaming.StreamJob: To kill this job, run:
12/12/28 12:35:20 INFO streaming.StreamJob: C:\Hadoop\hadoop-1.1.0-SNAPSHOT/bin/hadoop job -Dmapred.job.tracker=localhost:50300 -kill job_201212271510_0010
12/12/28 12:35:20 INFO streaming.StreamJob: Tracking URL: http://127.0.0.1:50030/jobdetails.jsp?jobid=job_201212271510_0010
12/12/28 12:35:21 INFO streaming.StreamJob: map 0% reduce 0%
12/12/28 12:35:38 INFO streaming.StreamJob: map 100% reduce 0%
12/12/28 12:35:50 INFO streaming.StreamJob: map 100% reduce 100%
12/12/28 12:35:56 INFO streaming.StreamJob: Job complete: job_201212271510_0010
12/12/28 12:35:56 INFO streaming.StreamJob: Output: output/Temp
4. Using the web interface or the JavaScript console, go to the output folder and view the part-00000 file to see the output (should match your testing results).
https://prologika.com/wp-content/uploads/2016/01/logo.png00Prologika - Teo Lachevhttps://prologika.com/wp-content/uploads/2016/01/logo.pngPrologika - Teo Lachev2012-12-28 20:18:002016-02-16 06:05:19Programming MapReduce Jobs with HDInsight Server for Windows
As another year is winding down, it’s time to review and plan ahead. 2012 was a great year for both Prologika and BI. On the business side of things, we achieved Microsoft Gold BI and Silver Data Platform competencies. We added new customers and consultants. We completed several important projects with Microsoft acknowledging two of them.
2012 was an eventful year for Microsoft BI. SQL Server 2012 was released in March. It added important BI enhancements, including Power View, PowerPivot v2, Reporting Services End-User Alerting, Analysis Services in Tabular mode, Data Quality Services, Integration Services enhancements, MDS Add-In for Excel, Reporting in the Cloud, and self-service BI for Big Data with the Excel Hive add-in. The next BI wave came with Office 2013 and added important organizational and self-service BI features, including PowerPivot Integration in Excel 2013, Power View Integration in Excel 2013, Excel updatable web reports in SharePoint, productivity enhancements (Flash Fill, Quick Explore, Quick Analysis, and so on), PerformancePoint theming support and enhanced filtering, better mobile BI support, and self-service BI in Office 365.
Microsoft added support for Big Data and Hadoop both on cloud with HDInsight Service and on-premises with the CTP release of HDInsight Server. Finally, we got the public prerelease bits of DAXMD to connect Power View to multidimensional cubes.
As we witnessed, BI is a very important part of the Microsoft data strategy. Although overwhelming in times, I hope the trend will continue in 2013 and beyond. In the spirit of the season, here is my top 5 wish list:
Continuing focus on integration and simplification – Fast-paced in nature, Microsoft BI has grown in complexity and redundancy through evolution and product acquisitions. Personally, I’d like to see further unification of the Multidimensional and Tabular models, so BI pros don’t have to choose which path to take and what compromise to make. Integration opportunities exist in other areas, such DQS and MDS, as well Tabular native support in the client tools (Excel and SSRS).
Extending mobile BI reach – Customers are asking for it.
Easier and simpler self-service BI – Excel 2013 has started the path but I think we can do a better job to simplify the user experience and compete more successfully with other self-service BI vendors.
BI in the cloud – This will enable interesting scenarios and extend the reach of BI products and services.
Enterprise lineage and change impact analysis – I think it’s about time to have this.
Most importantly, I hope to see Microsoft B having a renewed focus on customers in 2013. We should be listening more to our customers as sometimes as geeks we tend to be too much caught up in technology and we learn our lesson the hard way.
