fastverse / fastverse Goto Github PK

Hello, I wanted to suggest packages additions and removals

• nanotime is the first package to handle nanos in R, I think it should be used in replacement of clock
• clustermq is a package that leverage zeromq to send function to workers on a grid with 0 files involved
• qs is a binary format like fst that supports all objects (vs data.frames for fst) though no random access
• lubridate is (I think) fairly slow so I am surprised it’s on the list

Many thanks for this universe, I did not know kit, that's great to discover new packages
Regards

👋🏽 I maintain the cran checks badges. Please change to the new cran checks badge URL (e.g., https://badges.cranchecks.info/worst/dplyr.svg). Old badges at (e.g. https://cranchecks.info/badges/worst/dplyr) will be unavailable as of Jan 1st 2023.

• Blog post from June 2022 announcing the change: https://recology.info/2022/06/cran-checks-badges/
• New badges repo: https://github.com/r-hub/cchecksbadges
• Old badges repo: https://github.com/sckott/cchecksapi

I suggest that JuliaCall and alike should be removed. Based on my and many others' experience, Julia has failed to deliver the performance claimed by its enthusiasts. See two example posts:
https://discourse.julialang.org/t/with-missings-julia-is-slower-than-r/11838
https://discourse.julialang.org/t/my-experience-as-a-julia-and-r-user/83613

The misinformation and marketing of Julia have caused many R users, including myself, to lose valuable time. Think about this, if Julia is really that fast, why are there no major firms using it at all in the area of ML, CFD, or HPC?

Currently, more of Julia's issues have emerged. For example, see this post:
https://yuri.is/not-julia/

Hi,

First off, thank you very much for the package -- I am a big fan of the overarching goal. I just wanted to flag an error that just cropped up. As far as I can tell, nothing changed on my end. Removing and reinstalling the package (either with install.packages() or from a tar ball with R CMD INSTALL didn't resolve the issue. It seems like it stems from the rule() call within fastverse_attach(), but I couldn't resolve the issue. Any help with this would be greatly appreciated.

Minimal example illustrating the issue (in a fresh R session) with the incremental output:

> library(fastverse)
-- Attaching packages --------- fastverse 0.2.4 --
v data.table  1.14.3          v collapse    1.8.7      
v magrittr    2.0.3.9000      v matrixStats 0.62.0.9000
v kit         0.0.12          v fst         0.9.8      
-- Conflicts ------------ fastverse_conflicts() --
x collapse::fdroplevels() masks data.table::fdroplevels()
> fastverse_extend("ggplot2")
Error in rep("-", w) : invalid 'times' argument
> 

Nothing above is unique to ggplot2 -- same happens when it is a vector of packages, or when the package name is unquoted. And output of sessionInfo():

> print(sessionInfo())
R version 4.2.1 (2022-06-23)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Catalina 10.15.7

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils    
[5] datasets  methods   base     

other attached packages:
[1] fst_0.9.8              
[2] matrixStats_0.62.0-9000
[3] collapse_1.8.7         
[4] kit_0.0.12             
[5] magrittr_2.0.3.9000    
[6] data.table_1.14.3      
[7] fastverse_0.2.4        

loaded via a namespace (and not attached):
[1] compiler_4.2.1 parallel_4.2.1 tools_4.2.1   
[4] Rcpp_1.0.9     fstcore_0.9.12
> 

EDIT This issue only seems to arise in RStudio -- the same error does not arise when running R in console. RStudio version information below:

RStudio 2022.07.1+554 "Spotted Wakerobin" Release (7872775ebddc40635780ca1ed238934c3345c5de, 2022-07-22) for macOS
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) QtWebEngine/5.12.10 Chrome/69.0.3497.128 Safari/537.36

You could remove the dependency on magrittr and use the base pipe |> - it would require some fairly trivial rewriting of examples:

DT %>% gby(country) %>% gv(9:13) %>% fmean

needs some brackets for the fmean:

DT |> gby(country) |> gv(9:13) |> fmean()

I guess any examples that use piping to dot to send the pipe to non-first args will need changing to use underscore, for R version 4.2.0 and later...

Base pipe should be faster, but maybe you've already thought about this and decided to stay with magrittr?

https://paleolimbot.github.io/geos/index.html

To go with the other Spatial (SP) packages. Wraps the underlying GEOS C API and only has two recursive dependencies (libgeos and wk).

I have some examples showing integration with data.table and some benchmarks here: https://grantmcdermott.com/fast-geospatial-datatable-geos/#speeding-things-up-with-geos