Talk:DuckDB

Sources

@IgelRM: DuckDB appears to be fairly popular, so I have added a couple of independent sources that back up the claims. They are not excellent, but appear to be good enough for WP:GNG. I could simply remove the PROD hatnote, but would prefer you to take a second look at the article now. For the avoidance of doubt, I am 100% not connected to the project (learned about it by looking at the entry as a part of the WP:NPP). Sincerely, --Викидим (talk) 01:29, 23 March 2024 (UTC)[reply]

@Викидим: Thanks for reaching out and trying to improve the article. Unfortunately, popularity doesn't necessarily give notability and the sources you added ballistically say researchers introduced new database, which is not WP:SIGCOV. Regards IgelRM (talk) 01:38, 23 March 2024 (UTC)[reply]

We could redirect to Centrum Wiskunde & Informatica if you prefer as an WP:ATD. IgelRM (talk) 01:41, 23 March 2024 (UTC)[reply]

As I have stated, I am not vested in the outcome at all (I now know about DuckDB myself, this alone justifies for me spending 10 minutes of my time on editing). So feel free to proceed the way you like, including letting the PROD to run its course. Викидим (talk) 01:51, 23 March 2024 (UTC)[reply]

Hey folks. I've just removed the PROD tag after having expanded the article using reliable sources that I believe demonstrate WP:SIGCOV. Please let me know if I've done anything incorrectly. Jonathan Deamer (talk) 06:54, 23 March 2024 (UTC)[reply]

@Jonathan Deamer: solely several theregister refs is one source and the word reliable in the same sentence is rather novel. Also MotherDuck appears to be separate from DuckDB Labs, so maybe WP:HATSTAND. IgelRM (talk) 14:16, 23 March 2024 (UTC)[reply]

@IgelRM Agree that several from The Register is not as good as several from several sources, but it is noted as "considered generally reliable for technology-related articles" at Wikipedia:Reliable sources/Perennial sources. Understood that the MotherDuck reference may not confer notability for DuckDB, but I do think it's encyclopaedic to note this well-funded use of DuckDB. Jonathan Deamer (talk) 14:40, 23 March 2024 (UTC)[reply]

Yeah, I see. Sorry my reply came out a bit harsh. I just wanted to critique the sources, but it was of course fine to remove the PROD. IgelRM (talk) 21:47, 25 March 2024 (UTC)[reply]

I work at DuckDB Labs, which is an obvious conflict of interest, so I will refrain from editing the article. I would merely like to provide a few pointers for three books about DuckDB that are under publication:

DuckDB in Action - https://github.com/duckdb-in-action/examples
Getting Started with DuckDB - https://www.packtpub.com/product/getting-started-with-duckdb/9781803241005
DuckDB: Up and Running - https://www.oreilly.com/library/view/duckdb-up-and/9781098159689/

The authors of these books are not affiliated with DuckDB Labs. Szarnyasg (talk) 19:57, 25 March 2024 (UTC)[reply]

DuckDB is not an RDBMS

First sentence states:

> "duckDB is an open-source column-oriented relational database management system"

The reference used to back that statement states that duckDB is

> an analytical embeddable database system (page 611)

DuckDB website states that

> DuckDB is a fast analytical database

There is no support to back the claim that duckDB is an relational database management system. To my knowledge, duckdb is not even a relational database.

I suggest we phrase duckDB as an analytical embeddable database system, as the original source refers it to - at least the sentence is consistent with the primary source. Jorgecarleitao (talk) 04:52, 23 May 2024 (UTC)[reply]

Hey Jorgecarleitao!

I work at DuckDB Labs, It's stated on the Why DuckDB page: https://duckdb.org/why_duckdb

> To start with, DuckDB is a relational (table-oriented) DBMS that supports the Structured Query Language (SQL). Samansmink (talk) 11:07, 31 May 2024 (UTC)[reply]

As noted above, the DuckDB labs documentation states that "DuckDB is a relational database management system (RDBMS)." https://duckdb.org/docs/sql/introduction.html. I will go ahead and add this as a cite to the relevant text in the article. Of course welcome any further improvements. Plaur782 (talk) 20:33, 20 November 2024 (UTC)[reply]

DuckDB poorly match enterprise data storage

How to interpret statement below in the wiki: "…but match poorly the requirements of the enterprise data storage"

1) Isn't DuckDB using parquet file format with compression?

2) Isn't storing large amount of data in compression to save storage and yet optimized for data access are what enterprises need? Tom177y (talk) 12:20, 29 October 2024 (UTC)[reply]

Reading the cited source, the author of the cited book 'Research Software Engineering: A Guide to the Open Source Ecosystem' indicates "DuckDB is serverless and allows accessing Parquet files via a very fast SQL interface. This makes DuckDB a great tool for interactive analysis and transfer of large result sets, but is not so suitable for enterprise data warehousing." With that full context, the statement that 'but match poorly the requirements of the enterprise data storage' appears more of a defensible position, but presumably still the authors opinion based on their experience. I will go ahead and try to revised the text in the entry to better reflect what the author says in the source text. There may be standards for how to reference comments of this nature, and certainly open to further improvements. Plaur782 (talk) 20:19, 20 November 2024 (UTC)[reply]