DuckDB was added to the wall of fame of CWI, next to Dijkstra’s shortest path algorithm, the Atlantic Crossing of the Internet & the creation of Python
On the occasion of our 80th (really!) bday
Congrats to @hannes.muehleisen.org & @markraasveldt.bsky.social
@duckdb.org
@cwi-amsterdam.bsky.social
Nice. I think DuckDB helps with that. DuckLake if you want full-blown data lake. hehe
For organizing data, here is @paulgp.com example use case. github.com/paulgp/ipeds..., albeit saved as .duckdb format instead of .parquet.
Haha. I'm sure many read the replication code.
Hi James, I came here after seeing this. Sorry for the bombardment. x.com/joefrancis50...
Addendum: if storage is the concern go for parquet.
The next big thing of DuckDB Labs is ducklake and parquet sits at the storage layer (aka data files/blobs).
ducklake.select
As a rdatatable person, when I learned abt duckdb, wooow. The duckdb-cli is awesome for interactive poking. duckdb -ui if you need ide-like exploration.
1. A short vid: duckdb.org/2025/03/12/d...
2. Rdatatable fares well in the benchmarks, ofc duckdb is the king. duckdblabs.github.io/db-benchmark/
Parquet is the goat, in my experience. @hannes.muehleisen.org (ceo of duckdb) can't talk enough about it.
So, it is for anything you're storing, sharing, or querying.
parquet.apache.org
Feather is mostly for interchange such as R <-> Python.
arrow.apache.org/docs/python/...
Damn, the Dutch driving theory exam isn’t easy.
Thankfully, you can take it in English.
Damn snorfiets. 😂
Don’t underestimate it, prepare properly.
I recommend theorieexamen.nl if you plan taking the exam.
DuckDB output showing the number of items per country. Another DuckDB statement showing the total number of items.
Damn, 5 mil+!
#datagen
Hi #econtwitter #rstats,
Looking for great blog posts or talks on reproducible workflows w/ #QuartoPub in R/Python.
Also, examples of high-profile irreproducibility blunders 😉. I'm interested in literate programming, version control w/ Git/GitHub, & dep management with uv & renv. Any suggestions?
Claude Code one-shot this: tex → toml → typst!
🤯 Times have changed!!
Interesting finds: <SQLite JSON Superpower: Virtual Columns + Indexing>
- Store JSON document raw
- Create virtual generated columns using json_extract
- Add indexes to those generated columns
- Query JSON at full speed
dbpro.app/blog/sqlit...
#sqlite #interestingfinds
🙄
In one weekend, if your mind is in a good state, you can learn a lot
jq, lazygit, ripgrep, toml, tmux, and rely on tldr: github.com/tldr-page...
No one can keep up with the speed of the agents (agent mode), not me. I'm in due to fomo, and the autocomplete in Zed is not so good with github copilot.
Joining late to the @cursor party.
I participated in #veluweloop2025 with a wonderful team (# 116). 🚲 8.1 km and 🏃➡️ 11.8 km.
Does not mean we 7 years younger. 😂
Aha, so thoughtful! Thank you!
We received 2018 yesterday.
You don't check Hacker News (news.ycombinator.com/) daily, do you?
As for me, I rely on Hacker Newsletter, a weekly summary of the best of hn. buttondown.com/hacke...
If you want daily, they have Tech Daily: buttondown.com/tech-...
You did good indeed for pointing to that thread.🙏
I didn’t even think that has that intention, lol. I don’t think I have criticized prof Lumley. Who am I to criticize? I was just interested in reshape and wanted to express my love for base r. Hoping also if the reshape function can be improved further by the core r team.
Thank you for the many contributions you made to R. I rem my excitement when you're talking at useR about implementing pipeOp, and soon it was.
I actually didn’t know it was written by Lumley. That’s long time ago.
Oh I have not seen that.
I'd work with the date functions because you'd catch any formatting issues along the way.
I also dread the R v. Python "holy war". I say embrace both, like RStudio did (sorry Posit). There is so much power in both.
Aha.
miguelhernan.org/wha...
