SAS Viya surpasses SAS 9 in data management
The recent release of SAS Data Engineering represents a recommitment by SAS to data management, an area of strong product capability in SAS 9 which now has a promising future with SAS Viya. As data management processes are often the bulk of SAS workloads, this will be welcomed by many customers looking to modernise their SAS platform and move to SAS Viya.
Although SAS Viya has been available since 2016, the initial focus of development was high-performance analytics and machine learning using the in-memory engine, CAS. Point-and-click data preparation features were included but these didn’t come anywhere near the level of SAS Enterprise Guide or SAS Data Integration Studio in SAS 9.
Since then, a huge investment has been made in developing data management functionality within SAS Viya to bring it up to and beyond the level of SAS 9. These features are not just relevant to an analyst preparing data but provide a complete offering for a data engineer developing robust data pipelines and managing high quality data outputs.
Here are just a few of the ways in which SAS Data Engineering on SAS Viya exceeds the capabilities of SAS 9.
1. Unified interface
SAS Studio in SAS Viya combines the capabilities of SAS Data Integration Studio and SAS Enterprise Guide in one interface. Data engineers and data analysts can collaborate more effectively as they are all using the same tool. It becomes easier for data analysts to assist in data engineering tasks and for data engineers to refactor code or flows that have been developed by data analysts.
2. Custom steps
All the most common tasks in SAS Data Integration Studio and SAS Enterprise Guide are now available as steps in SAS Studio but what if you want something else? It’s now possible to create your own custom step for almost any task and reuse it across your data flows. You can even use custom steps that have been shared by other users in GitHub.
3. Python
Want to use some Python in your SAS data flow? No problem! Just include a Python Program Step in your flow and open up the Python Code Editor. This comes in handy for certain data transformations where Python excels or if you already have some Python code that you want to reuse.
4. Parquet
SAS Viya supports the popular Parquet file format, enabling read and write access with a LIBNAME statement. You can now run SAS data steps or procedures directly accessing Parquet data on Amazon S3 and Google Cloud Storage object stores as well as Parquet files on a shared file system. Support for ADLS is expected later. The smaller file size of Parquet compared with the standard SAS7BDAT file format (up to 92% reduction) together with the lower cost of cloud storage can result in huge savings on storage.
5. DevOps
SAS 9 was born in the days of competing version control systems and manual processes for moving code and content between environments. All that has changed with the widespread adoption of Git and DevOps practices. SAS Viya has built-in Git integration to manage changes to programs and flows in a Git repository. Pipelines can be set up to perform automatic testing before deployment of the changes to production.
6. Managing data outputs
Well-engineered data can support multiple analytical use cases but only if analysts can easily find the data they need and be sure of what it contains. Otherwise, they will waste time extracting data again from source and applying their own transformations. SAS Information Catalog in SAS Viya enables users to search for data assets and browse metadata including data quality profiling. This can reduce the effort spent of data preparation and speed up the development of new business use cases.
Make the move to SAS Viya
SAS Viya now offers the data management capabilities that customers have been waiting for, and significant additional benefits. There are many resources available to help in making the move, including the SAS 9 Content Assessment and SAS Communities. Just get in touch with your SAS representative to find out more.
