Best of SQL — August 2025

  1. 1
    Article
    Avatar of jobsJobs·38w

    🚨How one tiny SQL query nearly killed our production system 🚨

    A production outage caused by a seemingly innocent `SELECT *` query demonstrates how database queries that work fine in development can fail catastrophically at scale. The query pulled 720MB of data from 60,000+ records, causing memory pressure, network bottlenecks, and connection timeouts. The post provides practical solutions including selecting specific columns, implementing pagination, caching hot data, using EXPLAIN plans, monitoring slow queries, and testing with production-scale data.

  2. 2
    Article
    Avatar of freecodecampfreeCodeCamp·41w

    How to Design Structured Database Systems Using SQL [Full Book]

    A comprehensive guide covering database fundamentals, from basic data concepts to designing structured database systems. Explores the DIKW pyramid (Data, Information, Knowledge, Wisdom), database management systems architecture, ACID properties, and various data models including relational databases, NoSQL systems, and Big Data approaches. Includes practical examples using PostgreSQL and covers essential topics like data types, domains, database design principles, and the differences between structured, semi-structured, and unstructured data.

  3. 3
    Article
    Avatar of laraveldevLaravel Dev·42w

    You should stop using booleans in your database

    Argues against using boolean fields in databases, suggesting datetime fields as a better alternative for tracking state changes. The post acknowledges this is a controversial stance and notes that some truly binary conditions may still require boolean representation, but emphasizes that many boolean use cases can be better modeled with temporal data that captures when changes occurred.

  4. 4
    Video
    Avatar of fireshipFireship·41w

    DuckDB in 100 Seconds

    DuckDB is an open-source, embeddable SQL database optimized for analytical workloads through columnar storage. Unlike SQLite's row-based approach, DuckDB stores data column-wise, enabling faster aggregations, filters, and joins on large datasets. It features vectorized query execution, multi-threading, and can directly query CSV and Parquet files. The database excels at time series analysis and is already used by major companies like Meta, Google, and Airbnb.

  5. 5
    Article
    Avatar of bunBun·39w

    Bun v1.2.21

    Bun v1.2.21 introduces major features including unified SQL client support for MySQL, SQLite, and PostgreSQL through Bun.SQL, native YAML parsing, 500x faster string postMessage performance, and a native secrets manager for secure credential storage. The release adds executable compilation via Bun.build() API with cross-platform targets, Windows metadata embedding, and SIMD-accelerated ANSI escape removal. Package management improvements include security scanner API, enhanced audit filtering, optimized lockfile-only installs, and interactive update scrolling. The update fixes 69 issues with extensive Node.js compatibility improvements and bundler optimizations.

  6. 6
    Article
    Avatar of lonely_programmerLonely Programmer·39w

    SQL INNER JOIN – Connect What Matches!

  7. 7
    Article
    Avatar of singlestoreSingleStore·41w

    Scaling PostgreSQL vs. SingleStore: Overcoming Performance & Complexity Limits

    PostgreSQL faces significant scaling challenges as applications grow, requiring complex workarounds like read replicas, manual partitioning, and additional tools for caching and analytics. These solutions create operational overhead and architectural complexity. The database's row-based storage struggles with analytical workloads, while lock contention limits write throughput. AI workloads with vector embeddings further strain the system. SingleStore is presented as an alternative that combines transactional and analytical capabilities in a unified, horizontally scalable SQL engine with built-in vector search and hybrid storage formats.

  8. 8
    Article
    Avatar of jooqJOOQ·41w

    Consider using JSON arrays instead of objects to transfer data

    JSON arrays can provide 15-25% performance improvements over JSON objects for data serialization by enabling positional access instead of key-based lookups. While JSON objects offer better readability and debugging, arrays reduce overhead in both CPU processing and data transfer. The approach works particularly well in controlled environments where the client knows the exact structure and position of data fields, such as jOOQ's MULTISET operator implementation using SQL/JSON.

See all SQL archives