it's much easier to migrate EG projects to SAS studio instead of spend so much effort to migrate to Databricks, but still have a question - those sas code can run efficiently ? run faster than other tools? beside code change cost, running cost is a cost too
Please note that SAS Enterprise Guide includes dozens of sample data sets in the install, and SAS also ships with lots of sample data in SASHELP. However, the Orion Star data is provided with SAS courses and is tailored for each one, so it is not available as a download.
It sounds like we misunderstood your question so sorry about that! When you ask if the SAS code can run efficiently in EG- are you actually asking how fast does Enterprise Guide process your code vs SAS Studio? If that's not your question, please rephrase. Thanks!
Sas code is optimized for analytics in its own environment so typically will run much faster than data bricks and other spark engines which is typically used with python based code which although does the job nicely is not always optimized for the task at hand. There is a report available that does performance test against spark engines that reports significant performance gains in the Sas Viya analytical engine
Thank you for this video. Do the action workflows in Git also take care of replacing SAS9 legacy functions with Viya functions and can old paths such as in filename statements also be modified as part of this process? If not, won't you end up with a bunch of non functioning paths and code?
Hi and thanks for your questions. This particular migration from Enterprise Guide project flow to a SAS Studio flow converts the tasks from enterprise guide to similar steps in SAS Studio flows - so yes, they are replaced with what works in SAS Studio on SAS Viya. One of the task is converted to a step that contains working sas base code that performs the same as the original step do. Having said that, most sas 9 base code will work in SAS Viya as well - it is the same underlying compute engine and if there is a need to replace with another function that will be done. When it comes to paths - the example shown in the video doesn't have any references to local paths - all references were to libraries and in this particular case to libraries that pointed to sas datasets - so by using libraries we are able to abstract the references to local paths. At the moment, the migration process do not support any way of remapping paths in Enterprise Guide projects - so those would need to be remapped manually. One could add manual steps in github actions to support this for instance. For regular sas9 base code - there is a tool that comes with the content assessment package (that you can download) that provides a way to substitute local path references with other path references - that you would want to use on your target system. So for SAS base code you could use that.
