August, 2011

[Or: Rant - Code-Reviewing Random Code on the Internet.] I’m sure I’ve stumbled upon this post before once, and I hit it again today. I think it’s a little unfortunate this appears so high up in the results, as the post has a bunch of problems. I don’t want to bash on it too hard, as it did one or two things that are nice. One is using IDisposable as a substitute for C# lock() statement when using ReaderWriteLockSlim. This is good. Both lock() and using(IDisposable) are prettier...

Hello State Machine fans! [Update I changed the title to this post, the old one doesn't really work. New readers - if I lost you already, have you come across the WF4 State Machine introductory materials ?] Since there have now been several eras of support for state machines in WF4 and WF3 , I think it would be good to have information part historical stocktake, part making sure everyone is up-to-date (and a reminder to CTP users of CTP license terms), and so on, not so much for the regular readers...

No matter how much I read about WCF in online forums and blog posts, I still couldn’t grasp the whole idea behind the extensibility model until I was able to play with it myself. What I discovered along the way is how the entire channel model works and how the binding and endpoint address fit into the equation. Although this may seem like I’m diving in details quickly, I personally find it very useful to illustrate real world scenarios to really show you the power of the API. On that note, I’d like...

In my post for tabular projects on the Analysis Services team blog , I gave a high-level overview of what tabular modeling is all about. That post gives a basic overview of how to get started writing models to be deployed to the VertiPaq engine. But how do the tools work? And why is the modeling experience so different from multidimensional (OLAP/UDM) models? Read on. When we set out to write the pro tools, we had an interesting challenge. We needed to create a fully functional modeling tool inside...

One of our internal customers pinged me with a quick question today. He asked whether it is possible to set a default type argument for T if you have an activity of GenericType, say DataCollectorActivity<T>. I might have said this in passing by before but the answer is yes, the same technique is used for activities such as ForEach<T>. The workflow designer will look for the attribute in the MetadataStore first before resorting to reflection, so there are 2 ways you can do it: 1. Quick...

Tabular projects have a very simple structure. When you create a tabular project, you see two files in the Solution Explorer: TabularProject1.smproj and Model.bim. You are allowed only one .bim file per .smproj. Although you can sidecar whatever other files you like in your .smproj, such as text readme files, these other files cannot take a build action of Compile. Building a .smproj generates some more files. The build outputs go to the bin subfolder of the project folder. You get three files out...

Sometimes it happens that via the forums I learn about a new (to me) beast in the WF Zoo. Today that animal is WorkflowDataContext . The MSDN document is fairly terse: “Represents the data context of the current workflow environment and provides a bridge to bring workflow arguments and variables into the scope of data binding.” Putting that into everyday WF terms, how would you actually use one of these?   Application 1: Reading Variable and Argument Values from inside a CodeActivity  ...

Recently we've received a few customer feedbacks reporting WF4 workflow designer would have bad performance or hang in some cases even when the workflow is relatively small. We've worked with customers and found out that it was actually a WPF issue. Good news is, the WPF team has already release a hotfix for it. So if you are experiencing bad performance/hang with WF4 workflow designer, I would recommend you to install this hotfix and try again. Although I cannot guarantee it will solve the...

I have received a few questions from customers about how to build and deploy Analysis Services projects in an unattended way, for integration into an automated build process. The stock reply is a multi-step solution, where you build from command line and then deploy either by writing your own program (using AMO) or by calling the Deployment Wizard from the command prompt. It’s not exactly an elegant solution, but it works. Here is a slicker trick for tabular projects Denali. As I pointed out...

Introit - call it whatever, but today I’m recycling an old post. And I have one more little confession. When I started writing a new rehosted app today, I found out my old toolbox FAQ was fairly useless. It was even getting to the point where someone had to post this in the forum : “I found a number of posts online about how to populate a toolbox with the standard activities/icons. Unfortunately they are all wrong in one way or another. Some add internal non user tools, some get the wrong icons,...

We have two pieces of functionality in Denali that allow you to promote a self-service model in PowerPivot (v1 or v2) to a tabular model. One is i mport from PowerPivot , where you create a new tabular modeling project from a pre-existing PowerPivot workbook. Another is restore from PowerPivot (pretend this documentation exists, it will get there eventually), where you can restore a PowerPivot workbook in SSMS in exactly the same way as any other Analysis Services backup. The restore puts the metadata...

One of the benefits of using WF4 to create workflow is that it provides a graphical workflow designer. This means in most of the cases, developers don’t need to write code to define their workflows. Instead, they can use the graphical workflow designer we provided. This can not only dramatically improve the productivity but also bring down the barrier of authoring workflows so that even non-developers can do so. Another advantage of this kind of graphical design is that it’s intuitive and easy to...

After you import or restore from PowerPivot , you are left with a workbook with pivot tables and slicers to point to the PowerPivot model. The next logical step in the import/restore process is to re-point these objects to your new tabular model. How do you do that? Attempt 1: Change the PowerPivot data connection to point to the new tabular model. Problem: this connection is not editable. Try again. Attempt 2: Edit the data source for each pivot table to point to the new tabular model. Problem:...

A little while I begin a new rehosted designer project, and this time I want to get a little bit more creative with it, mixing in some of the stuff I have done earlier with XAML compilation. The more functionality like that I add, the closer to the Visual Studio scenario the scenarios it is capable of would become. And unlike the standard rehosted designer sample, Visual Studio hosts the Workflow Designer in its own app domain. There are a reasons for this, and also a few consequences of this design...

Sometimes people get an error message about “Access is Denied” when trying to import from a PowerPivot workbook and are perplexed by its cause. The user can open the PowerPivot workbook locally, but import fails with an error like this: Here is the root cause of this error. In my previous post about importing from PowerPivot , I explained that the workspace database server does an ImageLoad  and restores the ABF embedded in the PowerPivot workbook. This restore operation only succeeds if the...

As I mentioned in my previous post on deploying tabular models using msbuild , we now have AMO for PowerShell support. Unfortunately, the documentation for the AS PowerShell cmdlets didn’t make it into TechNet or into the product for CTP3, so they are quite difficult to use at this time (unless you are good at psychically divining functionality and parameters). This post should help fill the gap until RC0, when the docs come online. Once there are docs, you can use Get-Help <cmdlet name> to...

I’m always getting frustrated by little problems at work. One by one they are tiny but gradually they add up. Especially if they happen again and again every day! Running out of hard disk space on my C: drive on one of my machines is one of the ones which continually reoccurs. Usually it’s right in the middle of a productive moment, it breaks my concentration, and it fails a test I am running with a false result. Even after I delete the known obvious suspect wastes of space I still only get back...

Amir Netz’s two billion and seven rows demo at last year’s BI conference was not just demoware. Unlike PowerPivot, where the amount of data you can load is bound by the workbook size , the amount of data you can load into tabular models is limited only by server capacity. This is true at both design time and at deployment time. If you have a workspace database server that’s got enough RAM to hold and manipulate your billions of rows, you really can sort, filter, add columns, etc on your tabular model...