Hello!
Both, Python and R, have been around for quite a while now ... and have not replaced anything, yet ... but have complemented existing solutions. So my money is on that they won't replace SAS in the future.
SAS is an old trusted war horse, so you are probably on the save side to invest in learning it. And what might be more, statistical programming is a general craft that quite independent from the platform actually used. It is a way of thinking ... and approaching problems posed. 🙂
But of course there will always be new ideas and new tools in this rapidly changing world ... we have to keep pace with it as well as SAS has to. So, let's do our very best. 🙂
--FJa
Hello!
Both, Python and R, have been around for quite a while now ... and have not replaced anything, yet ... but have complemented existing solutions. So my money is on that they won't replace SAS in the future.
SAS is an old trusted war horse, so you are probably on the save side to invest in learning it. And what might be more, statistical programming is a general craft that quite independent from the platform actually used. It is a way of thinking ... and approaching problems posed. 🙂
But of course there will always be new ideas and new tools in this rapidly changing world ... we have to keep pace with it as well as SAS has to. So, let's do our very best. 🙂
--FJa
You are comparing apples and cucumbers.
Python is a programming language, with which you can do anything programmable, where modules for statistics are available.
R is a modern-day statistics (only!) environment.
SAS, OTOH, is a complete data warehousing framework. It provides metadata storage, access control, (long-time tested and certified) procedures for analysis, and a turing-complete language (DATA step) for ETL and data preparation. Since it's in one big package from one source, you also get exhaustive documentation (and a guarantee for function) from this one source.
Doing ETL (as an example) with R is a major PITA, so you need additional tools there. But you can integrate R into SAS through SAS/IML, which uses the same basic matrix model of data.
Addendum: there is so much SAS out there that SAS knowledge will be required, at least for legacy code.
Just like there is still a need for COBOL and PL/1 / mainframe programmers, even though both have been declared dead for decades.
Great explanation @Kurt_Bremser Thank you. I am kind of interested in finding out the performance differences in tasks that can be done in both i.e. where stuff overlap. Any thoughts?
Maxim 4. Try it.
And always include the time it takes to come to a solution in your calculations. That's where SAS shines, IMHO.
PROC Python? Don't knw that? Is that different from FCMP? Have you got a pointer for me by chance? ... as I could not spot it in the documentation.
Cheers
fja
edit:
Have it. Here for all other blinds: https://documentation.sas.com/doc/en/pgmsascdc/v_038/proc/p1iycdzbxw2787n178ysea5ghk6l.htm
🙂
SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.
Ready to level-up your skills? Choose your own adventure.