Watch this Ask the Expert session to see the SAS integration with DuckDB and explore how it empowers users to get faster, simpler insights from data.
Watch the Webinar
You will learn how:
- SAS/ACCESS to DuckDB eliminates the need for database setup.
- DuckDB’s in-process engine delivers exceptional speed for modern analytics workloads.
- Trends like self-service analytics, open data architectures and embedded SQL engines are reshaping the landscape – and how SAS is leading the way.
The questions from the Q&A segment held at the end of the webinar are listed below and the slides from the webinar are attached.
Q&A
Does Viya allow to download and load DuckDB extension(s) be it core or community?
Yes, we are looking to include the key ones by default for dark sites, but in general they are downloaded as required.
Are you likely to integrate with MotherDuck in the future?
It's not actually on the road map. However, I have done some testing myself with that and it seemed to operate perfectly well just through either ODBC or JDBC and that kind of method.
So, I have done some tests, but it's not specifically on the road map. If that is something people are interested, then I encourage you to raise speech requests for that and we'll look at that in more depth.
How do you specify specific DuckDB schema(s) stored in a *.duckdb physical file?
Need to confirm, but use option Schema = in conjunction with database=
Why are open file formats and columnar query engines critical for modern data ecosystems?
There are a couple of reasons. First, a columnar file format is self-describing, providing enough metadata so that a query or computational engine can access just the right parts of the data without unnecessarily reading the entire data set. Second, it allows for parallel operations, engaging all cores to work on the data simultaneously. In a modern data stack, these systems are deployed on cloud infrastructures distributed across various data centres. It is important to be mindful of how much data you are moving and how to maximise the value from infrastructure, especially since costs are often based on a consumption model rather than capital expenditure. Columnar file formats in a modern data stack are designed to help you extract the maximum return on your infrastructure investment, making them much more efficient.
How does SAS/ACCESS Interface to DuckDB simplify analytics for technical and business users?
Essentially, it's an Access engine that supports a large number of different file formats. Anand has spoken a lot about Parquet, but of course, there are also JSON and CSV formats—it's essentially a single place to support all those elements. It's truly open and supports the entire ecosystem. You can use implicit SQL, where we'll convert the SAS syntax into DuckDB syntax, or for broader self-service capabilities, explicit SQL is your friend, allowing you to use the entire DuckDB framework within. So, it's a simple place to start with powerful performance.
What competitive advantages does SAS Viya offer in an open, evolving data landscape?
It provides a single unified platform that brings together a wide range of analytics for practitioners, eliminating the need to switch between multiple ecosystems to accomplish their tasks. This enables the creation of highly complex workflows in a streamlined environment. From an organisational perspective, it offers the governance and security expected from a platform like this. Essentially, SAS Viya serves as a melting pot—delivering a governed platform where you can access diverse data sources and perform advanced analytics.
Does it support Authentication for AWS and Azure?
It does. It integrates very well with IAM and Azure Intra.
You mentioned S3 storage. Does it support Azure data lake storage?
It does partially today, but a complete Azure Data Lake access is part of the plan development.
Does SAS Viya support downloading and loading DuckDB extensions, whether they are core or community extensions?
Certainly, that's a straightforward answer. Yes, we are considering including some of the key extensions so they don't need to be downloaded directly from the DuckDB website, especially for dark sites without Internet access. We will provide alternate methods, but the short answer is yes. Some of this may require explicit SQL, but the answer remains yes.
R has DuckPlyr (https://duckplyr.tidyverse.org/), would that be sorted with the R interface of Viya?
Access Interface to DuckDB delivers a libname enabling users to use SAS code to access data via duckdb. This has nothing specific to R.
Does Proc DS2/FedSQL support DuckDB library?
Not at this time, but it is on the roadmap and should be available soon.
Are SAS R&D looking to reduce SAS V9 cloud cost of operation?
Nothing related to DuckDB integration which is SAS Viya 4 only.
While creating AI agents, we often use Redis for volatile and persistent memory. You previously explained how volatile memory can be handled using in-memory DuckDB. Is there any way to store permanent (persistent) data using in-memory DuckDB?
Yes, a table created in-memory can be persisted to disk if desired by either attaching a file backed database and saving the in-memory table to it. Or exporting the table to parquet file.
In Agentic AI ecosystem, how we can be able to achieve persistence memory by using DuckDB
By operating DuckDB in a persistent mode, i.e. backed by a database file.
Recommended Resources
SAS Data Access Product Page - Data Access | SAS
DuckDB website: DuckDB.org
Futurum study HTML page: Unlock AI Productivity With SAS Viya | SAS
DM SAS Insights blog master-page: Data Management Insights | SAS
Related eBook: Decisions You Can Trust
Please see additional resources in the attached slide deck.
Want more tips? Be sure to subscribe to the Ask the Expert board to receive follow up Q&A, slides and recordings from other SAS Ask the Expert webinars.
