Announcing DuckDB 1.5.0

We are releasing DuckDB version 1.5.0, codenamed “Variegata”. This release comes with a friendly CLI (a new, more ergonomic command line client), support for the VARIANT type, a built-in GEOMETRY type, along with many other features and optimizations. The v1.4.0 LTS line (“Andium”) will keep receiving updates until its end-of-life in September 2026.

DuckDB
DuckDB 1.5 with spatial updates: DuckDB 1.5 arrives with significant updates for #geospatial: GEOMETRY becomes a built-in data type, storage switches to #WKB with shredding for better compression, #CRS awareness is now part of the type system, and more! Here is a rundown of the most relevant new features in #DuckDB 1.5.
https://spatialists.ch/posts/2026/03/22-duckdb-15-with-spatial-updates/ #GIS #GISchat #geospatial #SwissGIS
DuckDB 1.5 with spatial updates – Spatialists – geospatial news

DuckDB 1.5 arrives with significant updates for #geospatial: GEOMETRY becomes a built-in data type, storage switches to #WKB with shredding for better compression, #CRS awareness is now part of the type system, and more! Here is a rundown of the most relevant new features in #DuckDB 1.5.

Spatialists – geospatial news
🌕 Hacker News 完整數據存檔:Hugging Face 開源數據集詳解
➤ 實時更新的技術社羣數位史料庫
https://huggingface.co/datasets/open-index/hacker-news
本文介紹了託管於 Hugging Face 的 Hacker News 完整數據存檔專案。該數據集收錄了自 2006 年至今的所有論壇內容,包含貼文、評論及各類交互數據,總數超過 4,700 萬筆。透過自動化的管道,系統每 5 分鐘同步一次最新數據,確保開發者能獲取即時資訊。資料以 Parquet 格式儲存,並按年月分層管理,極大提升了查詢與分析的效率,是研究科技趨勢與討論熱度的極佳數據源。
+ 終於有一個能夠直接透過 SQL 查詢的 Hacker News 數據集了,對於想分析多年技術趨勢的工程師來說非常實用。
+ 每 5 分鐘同步一次的架構很穩健,透過 Parquet 分層儲存的作法,既節省了儲存成本,也讓後續的數據挖掘變得輕鬆許多。
#資料工程 #開源數據集 #Hacker News #Python #DuckDB
open-index/hacker-news · Datasets at Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

¡Pruebas brutales con DuckDB! El MacBook Neo (el más barato de Apple) rivaliza con servidores en la nube hasta 4x más RAM en cargas de bases de datos pesadas. En cold runs, ¡hasta 2.8x más rápido que instancias cloud! Big Data en tu mochila 💻🚀 #MacBookNeo #DuckDB #AppleSilicon

Ever tried filtering on COUNT() with WHERE and hit a confusing error?

The problem: WHERE runs before GROUP BY, so aggregate functions don't exist yet. That's where HAVING comes in. It filters after grouping.

WHERE = filter individual rows
HAVING = filter aggregates

Read more: https://jamalhansen.com/blog/having-filtering-grouped-results

#SQL #Python #DuckDB #DataScience #Programming

HAVING: Filtering Grouped Results

WHERE filters rows before grouping; HAVING filters after. Need "only cities with more than 10 customers"? That's HAVING.

Jamal Hansen
ICYMI, here’s my review of #DuckDB performance improvement from v1.0 to v1.5 (just released). https://oxbowresearch.com/p/how-much-faster-is-duckdb-15-vs-10
How Much Faster is DuckDB 1.5 vs 1.0? A lot

I benchmarked every DuckDB minor release from 1.0.0 through the 1.5.0 dev build on TPC-H, TPC-DS, ClickBench, and SSB. The results tell a clear story of DuckDB quickly improving.

Oxbow Research
refactor: Update to DuckDB 1.5.0. · 81830dee4a

sport - This is the database schema together with a small Flask application that I use to track and publish my mileages on bikes, running achievements and a bunch of other sport related things.

Codeberg.org
And #DuckDB joins the group of software that uses cURL in the back /cc @bagder

Today I'll be upgrading a couple of things to #DuckDB 1.5 and yet again, the folks at @duckdb provide a perfect way to adapt to a breaking change, here getting the order of x/y right:

- in 1.5 you can set geometry_always_xy = true to have xy in lat/long order, otherwise you get a warning
- in 2.0 not setting it, you get an error, but you can ofc use false to get <= 1.5 default behaviour
- in 2.1 make the new one the default

They did the same in 1.3 for lambda style.

https://duckdb.org/2026/03/09/announcing-duckdb-150#spatial-extension

Announcing DuckDB 1.5.0

We are releasing DuckDB version 1.5.0, codenamed “Variegata”. This release comes with a friendly CLI (a new, more ergonomic command line client), support for the VARIANT type, a built-in GEOMETRY type, along with many other features and optimizations. The v1.4.0 LTS line (“Andium”) will keep receiving updates until its end-of-life in September 2026.

DuckDB
Big Data on the Cheapest MacBook

How does the latest entry-level MacBook perform on database workloads? We benchmarked it using ClickBench and TPC-DS SF300. We found that it could complete both workloads, sometimes with surprisingly good results.

DuckDB