Pinned Tweet
Follow
Hannes M�hleisen
@hfmuehleisen
Hannes M�hleisen’s Tweets
1
13
67
Show this thread
All the data profiling and vizualizations are powered by which enables us to get lightning speed queries.
1
1
2
Show this thread
Junior Engineers: This is a zero-shot deep learning classifier running on our hand-assembled GPUs backed by a performant fully distributed feature store, we call it Vanaheim
Senior Engineers: Yeah it wrote a Naive Bayes CTE in SQL, duckdb seems to handle 10x our planned load
2
6
65
Show this thread
In Memoriam Prof. Dr. M.L. (Martin) Kersten (1953-2022) cwi.nl/news/2022/in-m, founder of our database research group at CWI in 1985. We mourn on the day of his funeral, proudly remembering his many contributions to database systems research.
2
6
27
DuckDB is now the first DBMS (afaik) to persist ART Indexes. Starting on the next release (i.e, current master branch), no more waiting for PK/FK rebuilds on reloads, or worrying about losing track of your created indexes :-).
4
15
80
Topics to follow
Sign up to get Tweets about the Topics you follow in your Home timeline.
Carousel
In Memoriam Prof. Dr. M.L. (Martin) Kersten (1953-2022), fellow of CWI and emeritus professor of computer science at the University of Amsterdam. Sadly, he passed away on July 6, 2022. cwi.nl/news/2022/in-m
36
45
Friday I presented the Database Architectures group for siks.nl (presentation here: cwi.nl/~boncz/siks202) on the day both and won the prestigious VIDI grant. A great day for data systems research in The Netherlands!
1
11
54
You can now install the dev version of the awesome #rstats package directly from r-universe! duckdb.r-universe.dev/ui#package:duc
13
40
Happy to announce that I was awarded a "#VIDI" grant by the Dutch Research Council to create a "Responsible Decentralized Data Architecture" on top of
22
13
126
Excited that our proposal for another Dagstuhl seminar on "Ensuring the Reliability and Robustness of Database Management Systems" was accepted! Thanks to , our reviewers, and, of course, the co-organizers , , and .
2
6
46
The recording of 's keynote "DuckDB Testing - Present and Future" is now available
14
38
2
10
28
I just discovered DuckDB (And its CLI) and I could not be more excited about it. Querying parquet and CSV files is easy as pie. duckdb.org/docs/api/cli
4
23
137
DuckDB 0.4.0 "Ferruginea" released with query cancellation on CTRL-C and parallelism for most queries
2
28
82
I can't think of a single reason to use Spark over duckdb honestly. Orders of magnitude lower complexity.
1
4
8
based on my tweets yesterday, just wrote up how I increased query performance 80x by using DuckDB instead of postgresql.
4
20
142
Show this thread
New blog post by Richard Wesley: Range Joins in DuckDB
duckdb.org/2022/05/27/iej
3
25
98
I will again repeat that DuckDB is pure magic and being able to seamlessly query pretty large CSVs (a few GB) with it absolutely rules.
Sample query:
SELECT * FROM ‘your_file.csv’
Its that easy folks.
13
39
418
DuckDB-Wasm is an in-process analytical SQL database for the browser. Creator, (André Kohn), joins us to talk about DuckDB-Wasm, WebAssembly, and more.
Apple: apple.co/3MtXduZ
Spotify: spoti.fi/3sJf8WB
Google: bit.ly/3FX02lS
#webdev
8
19
Been experimenting with Parquet, and for the analysis of these. And for the visualisation.
Absolutely loving how fast DuckDB is for crunching many millions of bus locations. Just wish it had spatial queries baked in so I can skip GeoPandas entirely!
Quote Tweet
Locations in London where buses regularly remain stationary for longer than 1 hour.
Alt: Finding bus depots and parking bays using positions from @busopendata
Show this thread
1
3
37
WOAH! Just ran a series of denormalizing join queries that took 4 hrs on a 16 Node ra3.16xlarge #Redshift cluster in a flock of Lambdas running in < 5 mins! That's the diff between $832 and $12! 💰💵⏱#awslambda #awsredshift #optimized #AWS
7
17
85
Show this thread
It was fun to have with us yesterday and hear about the DuckDB story, its internals, and the research challenges around it. It was also cool to see Hannes compiling the whole DuckDB in seconds:) Thanks for the exciting talk, Hannes!
2
2
26
Show this thread
6
45
130
Sometimes a tweet stands out that makes you think "wow, I must try that!" A quick test of with ONSPD in , on a very basic laptop, counts postcodes by UK Parliamentary constituency in 1.49ms! Thanks
Quote Tweet
The power of @DuckDB and @ApacheArrow:
"We can select 304,851 interesting rows from all 1,547,741,381 in the 10 year dataset in < 3 seconds on a laptop!"
Great demo! twitter.com/GarsBar35Plus/…
1
7
29
Show this thread
Ibis 3.0.0 brings some incredible new features including the ability to mix SQL with Ibis expressions and ibis compatibility with a backend.
Read about the most important changes in our latest blogpost ibis-project.org/docs/3.0.2/blo
24
53
V3.0.0 brings other awesome improvements like the new DuckDB backend along with many performance improvements.
check out the official release post by ibis-project.org/docs/3.0.2/blo
1
4
9
Show this thread
I think DuckDB is awesome and would like to pull it into pyscript as a basic capability
2
7
26
Holy 🦆uck! Tried out with and got some speedups on loading data and replacing some pandas filtering.
1 mil rows for a app loaded in 120 ms vs 30 s (raw pandas)
Next step: analyze months of data
✍🏻 Read More:
4
27
121
Show this thread
Jordan Tigani needs no introduction. He was part of the team that created #googlecloud #bigquery . He is now designing a new analytical offering. In this episode of It Depends, Jordan takes us down the path of why we need to rethink our approach to moder…
1
2
4
GIF
Quote Tweet
Announcing: Tad 0.10.0, the fast, free, cross-platform tabular data file viewer desktop app! Now powered by @duckdb, yielding major perf improvements for loading and exploring CSV, Parquet and DuckDb/SQLite db files. Download at tadviewer.com
2
6
18
excited to share an alpha version of Developer, a reactive EDA-centric SQL tool for exploring and transforming datasets. Something I've always wanted as a data person :) Powered by and ✨
try it & let me know what you think!
github.com/rilldata/rill-
1:16
12.4K views
13
55
280
Show this thread
Announcing: Tad 0.10.0, the fast, free, cross-platform tabular data file viewer desktop app! Now powered by , yielding major perf improvements for loading and exploring CSV, Parquet and DuckDb/SQLite db files. Download at tadviewer.com
5
39
178
