szarnyasg

7 ms·

31.
by szarnyasg 3y ago
I wrote a page about this: https://duckdb.org/docs/guides/performance/file_formats TL;DR: if your workload is aggregation-heavy, Parquet is pretty good. If you have many joins, the lack of fine-grained statis
32.
by szarnyasg 3y ago
DuckDB also does not (yet) support adding foreign keys to a table, see: https://duckdb.org/docs/sql/statements/alter_table#add--drop... One typical way to work around this limitation is to create a new table,
33.
by szarnyasg 3y ago
DuckDB runs fine on laptops, including x86 and ARM architectures. It runs on macOS, Windows and Linux. It's not yet available as a conda package but can be installed from conda-forge (or from PyPI via pip). Additionally, it also works
34.
by szarnyasg 3y ago
You can limit the number of threads used by DuckDB via `SET threads = x`, see https://duckdb.org/docs/configuration/pragmas#threads (Disclaimer: I work at DuckDB Labs)
35.
by szarnyasg 3y ago
Vendors are tuning their implementation (e.g. the schema, indexing, and queries used to implement the benchmark) and also their database system, but they cannot tune the benchmark workload itself. Apologies if this was ambiguous in my comme
36.
by szarnyasg 3y ago
Rigorous database benchmarking is indeed very difficult and time-consuming. I spent the last ~7 years working on benchmarks for graph processing systems in the Linked Data Benchmark Council (LDBC) [1], originally established in 2012 as an E
37.
Extensions for DuckDB-Wasm(duckdb.org)
6 points by szarnyasg 3y ago|0 comments
38.
by szarnyasg 3y ago
Disclaimer: DuckDB Labs employee here. What does your 'serious analytics' entail? Using the full text search extension's macros (stemming, using match_bm25, ...), running regexes, computing aggregates? Are you doing highly-se
39.
DuckDB's CSV Sniffer: Automatic Detection of Types and Dialects(duckdb.org)
15 points by szarnyasg 3y ago|0 comments
40.
Why would a Neo4j person be so fond of an embedded, relational database?(info.michael-simons.eu)
3 points by szarnyasg 3y ago|2 comments
41.
DuckDB's AsOf Joins: Fuzzy Temporal Lookups(duckdb.org)
14 points by szarnyasg 3y ago|4 comments
42.
Even Friendlier SQL with DuckDB(duckdb.org)
32 points by szarnyasg 3y ago|0 comments
43.
DuckDB ADBC – Zero-Copy Data Transfer via Arrow Database Connectivity(duckdb.org)
8 points by szarnyasg 3y ago|1 comments
44.
by szarnyasg 4y ago
It's also worth to mention the idea of WITH ITERATIVE. This is a proposed SQL syntax extension that allows formulating iterative computations. For many problems (including graph traversals), this makes queries easier to write and execu
45.
by szarnyasg 4y ago
Quine seems to be an interesting project but I'm not too familiar with it. Its main feature, evaluating complex multi-way join queries on incoming streaming data, exists in the relational world as the Materialize database which leverag
46.
by szarnyasg 4y ago
I designed and maintain several graph benchmarks in the Linked Data Benchmark Council, including workloads aimed for databases [1]. We make no restrictions on implementations, they can any query language like Cypher, SQL, etc. In our last b
47.
Ethical Concerns of Code Generation Through Artificial Intelligence(sinews.siam.org)
2 points by szarnyasg 4y ago|0 comments
48.
by szarnyasg 4y ago
Not this time, the largest data sets are going to be graph500-30 and datagen-sf10k-fb. They have 17B and 9.4B undirected edges, respectively.
49.
by szarnyasg 4y ago
In LDBC, we have a separate benchmark for graph algorithms: https://ldbcouncil.org/benchmarks/graphalytics/ This focuses on untyped, unattributed graphs and includes algorithms which are often formulated following
50.
by szarnyasg 4y ago
A plug: if you are looking for TPC-style application-level benchmarks for database systems, check out the LDBC Social Network Benchmark [1]. It has workloads for both OLTP and OLAP systems. We designed both of these to prevent many of the c
51.
New S3 Alternative: Meet Telnyx Storage(telnyx.com)
3 points by szarnyasg 4y ago|2 comments
52.
by szarnyasg 4y ago
Egress-free S3-compatible object storage from Telnyx
53.
by szarnyasg 4y ago
Does anyone have some experience regarding the bandwidth between Hetzner's old US location, Virginia, and the big cloud providers (AWS/GCP/Azure)? My use case requires lots of ingress into these clouds. I am currently using C
54.
by szarnyasg 4y ago
Maybe they meant this paper? Daniel Oppenheimer et al.: Fortune Favors the Bold (and the Italicized): Effects of Disfluency on Educational Outcomes (Proceedings of the Annual Meeting of the Cognitive Science Society, 2010), https:/&#x
55.
Coub will be closed on April 1, 2022. Don’t forget to download all your videos(coub.com)
7 points by szarnyasg 5y ago|3 comments
56.
by szarnyasg 5y ago
The News Feed Eradicator browser plugin (available for Chrome and Firefox) supports hiding the recommended videos on both the YouTube home page and in the sidebar next to videos. It did wonders for my focus.
57.
by szarnyasg 5y ago
Casey Neistat did something similar in 2012, although he returned something resembling a commercial video and not a blank canvas. > After arriving at an agreed-upon treatment with Nike for its latest commercial, director Casey Neistat li
58.
by szarnyasg 5y ago
Was it Mike Stonebraker's talk "My 10 Fears about the Future of the DBMS Field"? (2018 Donald B. Gillies Memorial Lecture, presented at the University of Illinois): https://www.youtube.com/watch?v=DJFKl_5JTnA
59.
by szarnyasg 6y ago
Keynote on SambaNova's hardware architecture at the CIDR 2021 conference: https://www.youtube.com/watch?v=iHhHHBuk3W4
60.
SambaNova raises $676M at a $5.1B valuation to double down on cloud-based AI SW(techcrunch.com)
1 points by szarnyasg 6y ago|1 comments

More ›