Lightnews — Scholar-powered news

Davis Vaughan @davisvaughan.bsky.social · 4d

It is not internal! It is something I've been iterating on to help with performance based PRs in vctrs and elsewhere. I have gotten enough questions about it that I should probably release it for real at some point, but feel free to use it github.com/DavisVaughan...

GitHub - DavisVaughan/cross: Run Functions Across Package Versions

Run Functions Across Package Versions. Contribute to DavisVaughan/cross development by creating an account on GitHub.

github.com

1 3 9

Reposted by Davis Vaughan

Mattan S. Ben-Shachar 🎗️🇮🇱🇺🇦 @mattansb.msbstats.info · 6d

We all know filter, mutate, summarize, but the #tidyverse has some advanced hidden jems!

Recently a client sent me some code they were using to look up values between 2 tables based on date-times.

I replaced ~100 lines of code* with a single inner_join(by = join_by(overlap())).

Andrew Heiss @andrew.heiss.phd · 6d

I just learned about the `truncate` argument in {lubridate} functions, which means NO MORE HACKY paste0(year, "-01-01") code to build dates when converting years to dates in #rstats

library(lubridate)

yr <- 2025

# This doesn't work
ymd(yr)
#> Warning message:
#> All formats failed to parse. No formats found.
#> [1] NA

#-------
# TIRED
#-------
# I've done this for *yeeeeears*
ymd(paste0(yr, "-01-01"))
#> [1] "2025-01-01"

#-------
# WIRED
#-------
# The truncated argument tells {lubridate} that it can
# ignore up to 2 (in this case) formats to look for.
# Ordinarily `ymd()` looks for three formats: a year,
# a month, and a day. `truncated = 2` means it can skip
# the month and day parts
ymd(yr, truncated = 2)
#> [1] "2025-01-01"

ymd(c("2025", "2025-10", "2025-10-02"), truncated = 2)
#> [1] "2025-01-01" "2025-10-01" "2025-10-02"

2 4 28

Davis Vaughan @davisvaughan.bsky.social · 6d

Getting inequality/rolling/overlap joins in dplyr is definitely one of my proudest accomplishments 😆 Glad it was useful!

1 6

Davis Vaughan @davisvaughan.bsky.social · 6d

2️⃣ I ALSO HAVE STRONG FEELINGS ABOUT THIS

1 8

Davis Vaughan @davisvaughan.bsky.social · 9d

Thanks!

1

Davis Vaughan @davisvaughan.bsky.social · 18d

artisanal code 🤌

5

Davis Vaughan @davisvaughan.bsky.social · 19d

So good to see you!! (And your family 😂)

1 2

Reposted by Davis Vaughan

Lucy D’Agostino McGowan @lucystats.bsky.social · 19d

Another delightful #positconf in the books! So lovely to see everyone 🌻 I miss it already!

Especially excited about:
🤖 @minecr.bsky.social’s aifeedr
🌬️ Making my code pretty like @davisvaughan.bsky.social’s with air
✨ @emilhvitfeldt.bsky.social’s editable quarto magic
📈 everything @kjhealy.co said

2 2 25

Davis Vaughan @davisvaughan.bsky.social · 19d

We actually tried so hard to get baristas for conf itself, but the hotel policy about bringing in third party vendors is so strict! Luckily that does not apply for dev day 😬

1

Davis Vaughan @davisvaughan.bsky.social · 19d

Time for tidyverse dev day!

Can’t wait to review some PRs and drink some coffee ☕️ #positconf2025

A menu of tidyverse themed coffee drinks

1 1 26

Reposted by Davis Vaughan

Vinicius Oike @viniciusoike.bsky.social · 21d

I've been using Air since switching to Positron and have been loving it. My only criticism is the styling of tibbles (it just breaks them). Apparently that has been solved

# fmt: tabular

coming soon

#PositConf

2 8

Davis Vaughan @davisvaughan.bsky.social · 20d

I pretty strongly agree

1 5

Davis Vaughan @davisvaughan.bsky.social · 20d

Working on it!

1 4

Davis Vaughan @davisvaughan.bsky.social · 20d

Confirmed to be a bug, sorry about that

2 3

Davis Vaughan @davisvaughan.bsky.social · 21d

We’ve actually secretly taken over the air R package too and may use it later on as well!

1

Davis Vaughan @davisvaughan.bsky.social · 21d

It’s actually a Rust crate not an R package! But we are indeed very excited about the upcoming tabular formatting!

1 2

Davis Vaughan @davisvaughan.bsky.social · 21d

Air is our incredibly fast R code formatter!

Come learn all about it at 2:40pm in Regency V in the Sparking Development Joy session!!

We also have hex stickers!

#positconf2025 #rstats

1 5 33

Davis Vaughan @davisvaughan.bsky.social · 25d

Can't wait to see everyone at posit::conf(2025)!!

Here's my annual "things to do" list for food, coffee, bakeries, and bars near conf!

#positconf2025 #rstats

docs.google.com/document/d/1...

atlanta-posit-conf-2025

posit::conf 2025 Hotel Coffee ⭐ I think this looks the most promising, there are other locations that look decent even if this one seems new and has few reviews and is inside another building 7 minu...

docs.google.com

2 10 39

Davis Vaughan @davisvaughan.bsky.social · Sep 4

Because it's just a markdown document, I had the idea that maybe you can tell an AI about the extrachecks list and have it analyze your project for you. It's not a guaranteed success, but it's probably a reasonable first pass? I'd still personally double check everything manually too.

2 1

Reposted by Davis Vaughan

Ella Kaye @ellakaye.co.uk · Sep 3

Once again, @davisvaughan.bsky.social's extrachecks have saved me from a likely CRAN rejection for an upcoming #RStats package submission. Thanks Davis! github.com/DavisVaughan...

GitHub - DavisVaughan/extrachecks

Contribute to DavisVaughan/extrachecks development by creating an account on GitHub.

github.com

3 9 40

Davis Vaughan @davisvaughan.bsky.social · Aug 28

That's so cool!

Davis Vaughan @davisvaughan.bsky.social · Aug 27

Regardless of all of that, we'd definitely like people to try it out - particularly with dev purrr and dev carrier as that fixes a number of things that @shikokuchuo.net improved on.

3

Davis Vaughan @davisvaughan.bsky.social · Aug 27

I think it's fine to use furrr if you want. Note purrr won't auto collect globals, and I think purrr's behavior is probably better, but it could make converting feel frustrating. future also has a whole host of backends that are supported. We are also still shaking out a few bugs on the purrr side.

2 1 8

Davis Vaughan @davisvaughan.bsky.social · Aug 21

Here's one more example with factors that I find to be very compelling evidence that `replace_when()` is way easier to reason about than `case_when()` in some scenarios where the input type is very important github.com/tidyverse/fo...

Feature Request: `fct_mutate()` - change values (not recoding level names) in factor vectors more conveniently · Issue #372 · tidyverse/forcats

I find it annoying to do mutations on factor columns while validating against existing levels and preserving the levels in the output. This is something forcats could help with maybe? pets <- tibbl...

github.com

2

Davis Vaughan @davisvaughan.bsky.social · Aug 21

did a somewhat heavy deadlift a few weeks ago and tweaked my back so badly i couldn't lift the toilet seat the next morning

1 1