SAS Aggregated Reviews

SAS just plain sucks. I’m 30 and use it daily at my job. I will probably leave this company purely because of forced SAS usage.
It’s crazy expensive. R and Python are awesome and free.
There is very little online support or community
There is no future for SAS. It’s a dying legacy product. I want to move ahead in my career and SAS is not the answer.
– There are so many issues with just getting it installed correctly, especially in a locked down enterprise environment. Much of the functionality is locked away in costly add-ons.
– The GUI and tools surrounding SAS just plain suck compared to stuff like RStudio.
– The SAS macro language is extremely obtuse and painful. Some of the legacy syntax the base language still supports is also pretty dumb.
– It feels like SAS isn’t innovating and keeping up in offering connectors to new databases, functions that write to xlsx without making a gobbled mess, plugins for github, etc. Trying to do automate and run these modern things in SAS’s world is asking for pain and workarounds that barely work.
source

SAS is hated by many young members of the statistics community because of the nature of its language. There are separate blocks where you must import data, manipulate it, process it, and the language is not so “plan English.” R is “plain English” in the respect that it is comparable to HTML in how easy it is to follow (if not easier), and there really is not much of a format to follow beyond making sure you aren’t mixing commands and stuff.
Edit- This is more significant than everything else on why many younger people prefer R: SAS costs money (a good amount of money, too), R doesn’t. I forgot to mention that R also has tons of free convenient packages and things that are easy to work with at no extra cost.
source

I don’t hate sas. That would be like hating ibm or microsoft. Now that they are over the hill and toothless, we can ignore them.
source

Lack of console is a biggie. It’s my biggest quip with SAS in day-to-day use.
source

SAS has some really powerful functionality but also some really obnoxious aspects. Largely that’s because it’s been around since punch cards, and as a matter of policy, the creators don’t make any changes to the language that would stop old code from working as intended. So basically at no point has anybody done anything to simplify or standardize the language.
Some keywords have to follow a forward-slash, or be enclosed in parentheses, or be delimited by commas, while others that play similar roles don’t. There’s no consistency on this stuff between procedures. Any time they want to overhaul something or implement new ideas, they have to graft it clumsily onto the base language. Someone invented object oriented programming? Better design an entire new sub-language. There are probably a half dozen totally different languages rattling around in there: the base SAS language, SAS Macro Language, SCL, graph template language, and I don’t even know what else. Not that you have to learn them all by any means, but it’s like any time you try to do something new, you’re diving down some huge rabbit hole.
For another example, at some point they came up with Output Delivery System, which would be a unified syntax for every kind of procedure output from datasets to graphics to tables that would be compatible with HTML/RTF/PDF/etc. Great. But of course they didn’t get rid of the old functionality, and what’s worse, some of the new dataset structures are just different enough from the old ones that people still find the old ones useful, and you wind up having to learn both. Some of these procedures already had this sort of redundancies. Any one of these issues is manageable, but the overall amount of clumsy bloat is just unreal.
source

For me the biggest thing is that you’re forced to do everything in SAS. I try to use the best tool available for the job and I often integrate with a lot of different technologies/stacks, some old, some new. Trying to plug SAS into that progress has been an absolute nightmare for me in the past, mostly because of licensing issues. SAS ALWAYS has a solution for the problem, it’ll just cost you a fortune. When trying to prototype something that may or may not have value, it’s hard to justify spending thousands of dollars on something you may only ever use once.
The other limitation is that you’re at the mercy of SAS’s release cycle or you need to rewrite everything yourself. Very little cutting edge work is being published as SAS code (if any) and that limits ones ability to leverage new methodologies or algorithms without putting in a done of work.
source

I hate SAS because of how much it costs. Depending on which packages you get, licenses could cost $100,000+ per year. This makes it a total nonstarter for startups I’ve been in, personal projects, or even some large companies that haven’t been using it already.
Note, I’ve never actually used SAS (and I’m not sure how to try it). I hate it like a club that won’t let me in.
source

I hate SAS for several reasons.
– It costs a stupid amount of money
– The programming capabilities exists but are very specific to SAS. Its structure is unique and doesn’t teach you general programming structures very well.
– The GUI. Maybe I am just an elitist or something but the esistance of the GUI just bothers me.
– Graphics are painful and not pretty. I switch between python and R so use all the different packages (ggplot2, matplotlib, seaborn,plotly) so many different options that generate amazing, pretty graphics.
– Maybe they have changed but interactive graphics I have never seen from SAS. I like generating graphs that have selectors and stuff like in Tableau so users can select different options to explore the data set.
source