I think animals that hibernate for much of the year are pretty cool because what do you mean you get to only experience your favorite season?
Remote
Eric R. Scott
@LeafyEricScott@hachyderm.io
mastodon 4.7.3Scientific Programmer & Educator | Data Engineer with grantwitness.org | Mentor for @Posit Academy | Ecologist | Tea geek
Posts are mine and do not represent my employers in any way
If you can't help but do a lil dance when you eat something tasty, I think we'll be friends.
Migrated from fosstodon.org so I can see all my friends again!
757 Followers
599 Following
38 Posts
Joined July 18, 2025
Website:
GitHub:
Grant Witness:
I recently leveled-up my ad-blocking game by hooking my #pihole up to a #tailscale network so I can have it blocking ads on all my devices even outside of my home network!
I used to assume that Tailcale would be scary to set up and only for someone with a lot more experience and time on their hands than me, but it was surprisingly easy. They even had a guide specifically for this use case:
https://tailscale.com/docs/solutions/block-ads-all-devices-anywhere-using-raspberry-pi
Replying to
@gerrymcgovern@mastodon.green the data center in Pima county isn't even built yet and they are already taking water that isn't theirs to take. https://www.tucsonagenda.com/p/project-blue-s-water-trickery
#Tucson
I cannot for the life of me figure out a good way to read in a dataset made of hundreds of CSVs with *mostly* the same columns (in 
). I wrote a blog post that is basically an extended reprex. Comments and solutions are welcome!
Replying to
@defuneste@fosstodon.org @cd_newton@hachyderm.io The duckplyr version does though!! It works!!!
I went to my local middle eastern market today, not remembering that it is Eid. A little like Trader Joes the day before Thanksgiving except everyone was smiling and in a good mood. Anyway, Eid Mubarak to all who celebrate
*Finally* got around to testing new {purrr} and {mirai} features on the #UniversityOfArizona #HPC and it is such a big step forward for researchers who would prefer not to leave their R comfort zone like myself. Here's a gist that you can run from an interactive session to launch SLURM jobs as workers and use them for parallel computing with `purrr::map()`: https://gist.github.com/Aariq/d52a0d861e5a2c90540548caa47021af
My wife is amazing at guessing movies my mom wants to see from extremely vague descriptions.
1. The one about a woman that won an award
2. The funny space one with the tree
She got both of these right, can you?
Replying to
@lunarloony@dosgame.club @DrHyde@fosstodon.org @mlanger@mastodon.world not to mention the air and noise pollution
Replying to
@CalmPresence @PetraPhoenix fully in support of growing things from the grocery store! Lychee seedling are super pretty, and I’ve got some sweet potato slips growing in a pot on my front porch!
Replying to
@geospacedman@mastodon.social @danwwilson@rstats.me I mean, sounds like that would work! "Cleaning" all the CSVs so that I/a user can use the tools that they're used to is definitely a possibility.
Replying to
@dodecadron@datasci.social @rdnielsen@floss.social @duckdb@mastodon.social Sweet! I'll try this out.
Replying to
@pkw I think that's what I'll need to do, but it seems like something arrow should be able to do (with or without the R bindings)
Replying to
@defuneste@fosstodon.org and if I could just specify columns and types I would! That's what I try to do multiple ways in the blog post. Sure seems like it should be possible!
Replying to
@rdnielsen@floss.social Hmm... yeah maybe a SQL or @duckdb@mastodon.social way to ingest all these CSVs into a database despite the differing number of columns.
What is the criterion that the Criterion Collection uses? Good movie?
Replying to
@dgkf@fosstodon.org I'm mostly joking. I think it's great to have something more integrated. Github actions was a huge game changer compared to Travis CI on GitHub, for example, so I imagine Forgejo Actions will be a similar kind of improvment. I do really like the woodpecker model of steps as containers that a volume gets mounted to successively though!
Replying to
@JoeSondow@dice.camp Marathon (the original trilogy) and Marathon (the new extraction shooter)
#marathongame
Replying to
@PetraPhoenix if your goal is to grow a tree that produces tasty fruit, this is not the way. Apples are not “true breeding”, meaning if you plant seeds from a Fuji, you won’t get a Fuji, rather you’ll more likely get something like a crabapple.
Replying to
@dgkf@fosstodon.org @defuneste@fosstodon.org OOh, if I had time this would be a fun project and a great excuse to learn more about codeberg and woodpecker CI.
Replying to on fosstodon.org
@dgkf@fosstodon.org But I just learned woodpecker CI (and I think I like it better)!
Replying to
@PetraPhoenix Basically all apple trees are grafted. Once you've bred a variety that's tasty (through cross pollination and growing from seed), you take cuttings and graft them onto rootstock, so all the Fuji apple trees are basically clones of the first Fuji apple tree. Even if they self-pollinate, there is genetic recombination and the seeds won't be like their parents.
Replying to
@geospacedman@mastodon.social Looks like the memory usage for the fread methods is much higher, so that's not going to help
Replying to
@cd_newton@hachyderm.io @defuneste@fosstodon.org duckdb::duckdb_read_csv() (the R fucnction) does not have the union_by_name argument 😭
Replying to
@patriksvensson I love this except for the missing apostrophe and missing period. I don't think I could deal with that
Replying to
@chiraag@mastodon.online Yeah...I also thought there would be a way to specify that a column is optional with {arrow} in R, but if there is, I haven't found it.
Replying to
@nxskok@cupoftea.social Yeah, this works (I have an example of this in the blog post), but it's not ideal because the memory overhead is MUCH greater than read_csv(fnames).
Replying to
@geospacedman@mastodon.social For this particular project, I think I need a database-like approach. I had high hopes for rearranging CSVs in hive style partitioning and using arrow::open_dataset(). Looks like DuckDB is going to be the way to go.
But in general, it would be nice for read_csv(files, col_select = 1:5) to "just work" even when some of the files have 6 columns, because it's a lot more performant than map(files, read_csv) |> list_rbind() (or the data.table equivalent, I'm guessing)