Transferring content from SAS Viya for Learners 3.5 to 4: Welcome to The Matrix.

2 Likes

There is a lot of excitement around our new SAS Viya platform for teaching and learning: SAS Viya for Learners 4. And for good reason – there are a ton of new upgrades to our software that will make teaching analytics in the classroom even easier. From streamlining of our SAS Visual Analytics interfaces, to PROC PYTHON in SAS Studio, to easier open-source and GitHub integration, and beyond, I’m certain that you’ll have a new list of personal favorites soon. And if you want to see what I’m particularly excited about right now, then read my article here: Need reasons to get excited about SAS Viya for Learners 4 (VFL4)? Start Here!

But, let me level with you – change can be scary. Perhaps you’re a long-term user of SAS Viya for Learners 3.5. And you’re worried about what will happen to all the stuff you’ve built out in that cloud. And let me frighten you a bit more by saying this:

None of your data, or work, will transfer automatically from SAS Viya for Learners 3.5 to SAS Viya for Learners 4. Why? Because they’re separate cloud deployments (VFL3.5 is in a SAS Cloud and VFL4 is in Microsoft Azure) and we cannot connect the two clouds.

Do I have your attention now? Did you heart race just a little bit? And did I just rain on your VFL4 parade?

My intention in this article isn’t to increase the amount of anxiety in your life. Far from it. Instead, I’m trying to prompt you towards action. Because there are ways to get some of your data + projects transferred from VFL3.5 and into VFL4. Moreover, there are ways to share across SAS Viya 4 environments that will (1) help professors assess student performance and (2) allow students to collaborate on projects.

But let’s start at the end to provide some global framing. With a conversion matrix, for the SAS Viya tools most commonly used by academics:

That Matrix is not like “The Matrix”. But it will hopefully help simplify the process a bit. That is, of course, after I unpack it a bit, with an act in three parts.

Part 1: Use GitHub to transfer code… or transfer it manually

Let me say it one more time: GitHub is a wonderful tool for collaboration and code sharing. And all programmers should learn it ASAP.

I believe in that statement so strongly that I even wrote two more articles leading with that. Both show you how to pull and push files to a GitHub account. And you can read ‘em here:

These two articles will get you well on your way towards transferring your code and files in both SAS Studio and Jupyter. Moreover, I’ll add something not found in my earlier post – and show you how easy it is to add a collaborator in a Git repo. Easy as in just 4 easy steps:

Go to your GitHub repo and find the Settings button. For example:

Next, find – and click, Collaborators on the left:

Confession: you’ll probably be prompted for a password. But I don’t want to include that as the 5^th step, because I promised 4 easy steps 😊 Instead, find Add people on the Manage access screen:

Finally, find an invite friends!

Yup – that easy.

Finally, the less ideal approach is to simply download your Jupyter Notebooks + files manually in Jupyter. And this can be done with a couple of clicks. Once in the environment, right click on the file(s) and select Download:

The files will then download to your local environment, where you can then upload the files into either (1) your new Jupyter environment or (2) your GitHub repo. This process isn’t ideal – but is simple and will work in VFL3.5, where GitHub was not directly integrated.

Part 2: XML code and VA Reports

What the heck is XML? Well, it’s Extensible Markup Language, of course! And it can be used to transfer your SAS Visual Analytics reports from one environment to another. Again, this is great if you’re migrating your VA reports from VFL3.5 to VFL4. Or, perhaps, you’re a professor working in VFL4 and want to see student’s homework also in VFL4. Or students collaborating on a single report in VFL4. Lots of permutations options here on sharing/migrating.

How do I make the magic of the XML transfer work? Well, there is already a SAS Communities article for that! And it can be found here: Sharing Reports in SAS Viya for Learners.

That’s what we call a brief, but informative, section!

Part 3: SAS Model Studio Projects

Ah – our beloved SAS Model Studio projects. For those getting acquainted with SAS Model Studio in SAS Viya for Learners 4, you might have noted the following options under the Additional actions for project menu:

What, what? Import? Of our projects! Yay! So, this MUST mean we can export.

And we can export. I’ll slide into another VFL4 environment where I’ve already built some SAS Model Studio projects and show you how:

I’ll focus on MortgageAnalysis and click on the Option button. I’ll then find Export:

You’ll get a message like:

And that sounds very important. When SAS Viya is done with the export, it’ll be in your download folder, as specified by your browser. And it will appears as a fun zip file:

And, class, what do I do with that ZIP file? All together now: upload it in the other environment as a zip file! So, I’ll do exactly that by, again, clicking Additional actions for project >> Import >> Data Mining and Machine Learning:

I then get this window:

The File (zip) part is easy. In my case it’s the MortgageAnalysis.zip file that I just showed you. For Data, it needs to be a CAS table in your existing environment. So, if it’s not yet there… you’re need to get it load. Mine is _PREPPED. And then you can click Import, once you have those required pieces:

Another friendly “I’m working on it” message appears:

And your project will be there when it’s ready:

Yay! I’ll open the project, just to prove I have a project with some pipelines:

Now you’ll likely notice – quickly – that not all the pipelines are transferred. For example, an interactive model pipeline won’t be transferred. And the primary reason is that Global metadata cannot be exported and imported among different environments even if they are the same version. I straight up stole that from @BethEbersole and this great post: Moving SAS Model Studio Custom Pipelines to a New Environment.

Beth also shows you how to import and export Custom Pipelines from SAS Viya 3.5 and import them into a new environment. Unfortunately, I’m not finding the current version of SAS VFL4, which was built on the 2024.03LTS version of our software, compatible with the VFL3.5 build from 2020. Instead, I get this:

That error aside, I’d love to hear if others have found a hack around to get individual pipelines transferred from VFL3.5 to 4. Like perhaps just the pipeline containing your champion model champion model 😊 And if you figure it out, I can add it to the information above.

Recap

Change can be scary. Especially when moving from one cloud to another and upgrading your software.

Hopefully, this article will help ease your transition. If there are still outstanding questions of issues, please know that we have an entire SAS Academic Programs team that can help you integrate analytics into your learning. So, please start a conversation below if you need support!

Happy learnin’! And if you haven’t switched to SAS Viya for Learners 4 yet, please check it out here!