name one reason why R sucks
> protip: you can't
>>60184700
Fuck you.
>>60184757
well memed m8 lol
>>60184700
(((((programming))))) language where basic branching has massive performance hit
>>60184700
It's not numpy
The idioms are a little "different" compared to modern languages. I have to re-run the swirl tutorials every time I haven't used R in a few months to remind myself how to do things the "R way". After learning Bash you can definitely see the UNIX influences on common R commands (paste, cut, source, ls, rm, grep, etc.)
>>60184700
It's one of the tools we use for prototyping data analytics/algorithms and its great. I prefer it to Python actually.
But it is basically built for and by statisticians. It's probably a bad language for any other purpose.
>>60184700
S L O W
L O W
O W
W
I use it to make graphs for management with datasets that make localc/Excel etc shit the bed
It's a bit idiosyncratic, but gets the job done
>>60184700
There is basically no reason to use it now thanks to the Python libraries for stats.
R fails when your dataset is larger than your RAM (unless you use something like microsofts R server) or when you need multi threading. Apart from those two limitations it's pretty good for analytics. PowerBI is cool too (I love dashboards).
>>60184700
2GB mem limit
>>60185138
cant-vectorize-fag detected
>>60185578
32-bit fag detected
>>60185613
IDK I havent used R for the past 3 years or so, do they have a 64bit version now?
>>60184700
It's not Fortran.
>>60185120
Pretty much this. For statisticians by statisticians. And it does that job pretty damn well.
>>60184700
Most good and bad points have been mentioned. One thing to add is that there are better and more consistent alternatives to much of the stuff in base. This is often confusing. Also packages vary widely on quality.
I work with larger data, and without data.table I couldn't get anything done, that is how shit base is.
Would anybody here recommend dplyr over over data.table?
It's bloated
not Julia
>>6018470
I love R. The only downsides I can think of are
It takes a while to figure out how to program stuff without using a for loop for everything
The base-syntax is not very intuitive. I hate that I didn't discover dplyr a few years earlier.
>default parameters in between required arguments are allowed
who thought this was a good idea?
>>60185595
Not him, but when i did the coursera courses I found it to be unreasonably slow at loading data from many files. Probably I wasn't good enough to figure out how to do it "the R way".
>>60190285
Nope. R can be very slow even with vectorization.
I don't really know it, but I worked a contract at a huge pharmaceutical company where they were doing protein modeling to find new drugs on a huge AWS cluster (like 2500 nodes) and R was involved.
R as in Rust?
Sorry I'm new here
>>60191250
No R the language called "R"
>>60191342
gotcha, thanks newfriend