#database

SQLite's strict tables enforce rigid typing, preventing type mismatches like inserting text into integer columns and rejecting bogus column types. Flexible typing is still available via the ANY type. Downsides include inability to alter existing tables to strict, requirement for SQLite 3.37.0+, and slight theoretical overhead.

A project called pgrust has rewritten PostgreSQL's query engine in Rust, now passing 100% of the standard Postgres regression tests, marking a significant milestone in the effort to port the database to Rust.

ClickHouse is winning the observability wars, emerging as the dominant database for logs, metrics, and traces. Its superior compression, query speed, and high-cardinality handling are driving adoption over Elasticsearch and TimescaleDB, with many modern observability tools now built on it.

The author describes creating EventSourcingDB, a new open-source event store database written in Go. After struggling with existing solutions like EventStoreDB and PostgreSQL for event sourcing, they built a lightweight, HTTP-native event store focused on simplicity and developer experience.

Multigres now supports PostgreSQL's LISTEN/NOTIFY feature across pooled database connections. This allows applications using connection pooling to receive real-time notifications without losing events, solving a longstanding compatibility issue between pooling and asynchronous notifications.

A DuckDB community extension lets users query Snowflake, Databricks, BigQuery, PostgreSQL, MySQL, and other systems via ADBC drivers. It supports read_adbc and ATTACH for running SQL statements as if the database were local. The open-source project is planned for contribution to the Arrow project.

This first part of a two-part series explains the theoretical foundation of B-tree indexes in PostgreSQL, covering their structure, how they store and organize data for efficient lookups, and the mechanics of search operations within the index.

The article discusses how database deadlocks can cause application downtime, explaining the mechanics of deadlocks in MySQL and PlanetScale's approach to handling them. It covers common causes, detection strategies, and best practices for minimizing their impact on production systems.

Xata introduces a new GitHub app that automatically creates a dedicated Postgres database branch for each pull request, enabling isolated testing and previews directly from the PR workflow.

sqlite-utils 4.0rc3 has been released with new support for compound foreign keys and case-insensitive column name matching following SQLite's convention. The stable 4.0 release was delayed as additional changes accumulated, including a breaking change to the table.foreign_keys interface.

PgBouncer is a lightweight connection pooler for PostgreSQL that reduces overhead by reusing connections. It works by maintaining a pool of database connections and distributing them to clients as needed, supporting session, transaction, and statement pooling modes to improve performance and scalability.

The article explains why PostgreSQL systems must use strict memory overcommit settings to prevent the Linux OOM Killer from terminating the database process. It details how memory overcommit allows PostgreSQL to allocate more memory than available, leading to out-of-memory conditions, and recommends setting vm.overcommit_memory to 2 to avoid such crashes.

ClickHouse is emerging as the dominant database for observability, outperforming Elasticsearch for logs, metrics, and traces due to its columnar design and compression. Major observability platforms like Grafana and Datadog now use or integrate with it.

The article argues that as modern data stacks rely on ORMs, query builders, and AI-generated queries, fewer developers and analysts actually read or write raw SQL, which leads to performance issues, security vulnerabilities, and a loss of deep understanding of how databases work.

Scalar has revamped its documentation for ScalarDB and ScalarDL, aiming to help users start faster, stay updated with changes, and navigate cleaner, more organized docs. The update includes improved structure, clearer guidance, and better version management for developers.

ClickHouse partnered with a major tech company to scale LogHouse, a high-volume logging system handling 1.59 quadrillion rows across AWS, GCP, and Azure. The system processes 1.2 PB of new data daily, using optimized compression, tiered storage, and real-time replication to achieve sub-second query performance at minimal cost.

ScyllaDB has introduced a new trie-based index that delivers up to three times more throughput compared to its previous indexing methods. The trie index reduces memory overhead and improves performance for both point lookups and range scans. This enhancement is part of ScyllaDB's ongoing efforts to optimize NoSQL database performance.

The article explains how co-locating workflow state with application data in Postgres enables reliable distributed systems by leveraging database transactions and atomicity. This approach simplifies coordination, making workflows more robust and consistent without relying on external state management.

SurrealDB has launched Scale, a new cloud tier designed for high availability and large-scale deployments. The offering includes multi-region support, horizontal scaling, and enhanced reliability features for enterprise workloads running on SurrealDB Cloud.

A proposal to fuse the database and programming language into a single system where the entire system state is treated as a program, eliminating the impedance mismatch between application code and data storage.

Pgblame is a new tool that helps identify which Vercel deployment caused a slowdown in your PostgreSQL database by correlating deploy timestamps with database performance metrics.

ClickHouse is emerging as the dominant database for observability workloads, outperforming alternatives like Elasticsearch in query speed, compression, and cost efficiency. Its columnar storage architecture makes it particularly well-suited for handling large volumes of logs, metrics, and traces in modern monitoring stacks.

The Stockholm BMA meeting for SQL:202y advanced new SQL features including enhanced temporal support, improved JSON, and property graph queries. Discussions covered polymorphic table functions and ratification of previous decisions toward the next standard revision.

PL/PHP is a procedural language handler for PostgreSQL that enables users to write database functions using the PHP programming language.

Radio-Browser.info provides a database of over 60,000 radio station streams from around the world, allowing users to search, browse, and access internet radio stations for free.

Supabase reported a platform incident where existing databases cannot be restarted or created, indicating ongoing service disruptions for users.

A Go developer created a new zero-syscall IPC mechanism called hft-ipc, achieving 18 million transactions per second to address slow database insert performance. The implementation uses a wait-free, lockless ring buffer design with zero allocation and no system calls during operation, inspired by LMAX Disruptor and DPDK.

NVIDIA GQE is an open-source library that helps developers build GPU-accelerated SQL query engines. It provides building blocks for query compilation, execution, and data management on GPUs, enabling higher throughput and lower latency for database and analytics workloads compared to CPU-only engines.

This paper presents a set-theoretic approach to detect logic bugs in DBMS inner-join optimizations, aiming to improve database reliability by identifying incorrect query optimizations.

The author announces they are leaving their current company to focus full-time on LadybugDB, a lightweight embedded database for Rust. They reflect on the journey so far and outline plans to accelerate development, improve documentation, and build a sustainable open-source project. The post also includes a call for community feedback and contributions as LadybugDB moves forward independently.