This article traces the evolution of data recovery from manual backup restoration through log replay to automated replica switching, explaining RPO/RTO trade-offs, and details how 10M QPS systems implement shard-level automatic rebuild with throttling, verification, and continuous recovery as a runtime capability.
This article explains why vector databases are general-purpose similarity search engines beyond RAG, covering embedding model selection (BGE), non-RAG use cases like recommendation and anomaly detection, core operations with ChromaDB and HNSW, selection criteria for ChromaDB, FAISS, Milvus, and Pinecone, and five practical engineering lessons on recall evaluation, API workarounds, distance vs similarity, chunking strategies, and metadata filtering.
This article explains the logical execution order of SQL clauses (FROM, WHERE, GROUP BY, HAVING, SELECT, ORDER BY, LIMIT), why SELECT aliases aren't visible in WHERE, how HAVING differs from WHERE, JOIN/ON placement, LEFT JOIN pitfalls, and how to verify with EXPLAIN, distinguishing logical vs. physical execution order.
This article explains why SQL NULL represents an unknown state rather than a value, how three-valued logic (TRUE, FALSE, UNKNOWN) breaks equality comparisons, and demonstrates practical pitfalls with WHERE clauses, NOT IN subqueries, aggregate functions, and MySQL's NULL-safe operator.
pg-jev is a PostgreSQL extension that wraps TypeSafe's Jev library into SQL functions, enabling semantic classification, probability scoring, and label selection directly in queries without vector indexes or embeddings, with performance benchmarks and operational caveats.
This article provides a structured decision framework for choosing databases in new projects, covering four key questions, relational and NoSQL comparisons, Xinchuang (domestic) database options, distributed NewSQL systems, a practical decision tree, typical technology stacks, and guidance on writing selection documents with POC validation.
This article provides a deep technical analysis of StarRocks' vectorized execution engine and cost-based optimizer, covering pipeline scheduling, vectorized operators, CodeGen/JIT compilation, statistics-driven cost modeling, parameter tuning strategies, and production troubleshooting methodologies for low-latency analytical queries.
The article explains why a MySQL SELECT * query scanning 20 million rows won't cause server-side OOM due to its streaming protocol using a 16KB net_buffer, but warns that full table scans can pollute the Buffer Pool, though InnoDB's midpoint insertion strategy and innodb_old_blocks_time (default 1s) protect hot data from eviction.
This article details a five-phase migration process for China's Xinchuang database replacements, covering Oracle-to-Dameng and MySQL-to-Kingbase routes, compatibility mode selection, schema conversion, data migration, application refactoring, regression verification, and critical dialect pitfalls like auto-increment, upsert, date functions, and empty-string handling.
This article explains how TiDB's NewSQL architecture eliminates manual sharding by providing horizontal scaling, ACID transactions, and MySQL compatibility through a distributed architecture with TiKV storage, PD scheduling, and TiFlash for HTAP workloads.
This article explains ten advanced SQL concepts including CTEs, recursive queries, temporary functions, pivoting with CASE WHEN, EXCEPT vs NOT IN, self-joins, ranking functions, delta calculations with LEAD/LAG, running totals, and datetime manipulation, with code examples for data science interview preparation.
This article explains the difference between SQL's written clause order and its logical execution order, detailing why WHERE cannot use column aliases, how HAVING filters groups, the impact of JOIN/ON placement in LEFT JOINs, and how to verify execution plans with EXPLAIN.
This article explores MySQL 9.7's key innovations including native VECTOR type for semantic search, JavaScript stored procedures for in-database logic, parallel bulk import from object storage, and security upgrades like mandatory caching_sha2_password, with practical code examples and upgrade guidance.
This article explains Oracle database backup and recovery fundamentals, covering failure types, logical vs physical backup, RMAN configuration, level 0/1 incremental backup scripts with cron scheduling, flashback technology, and complete vs incomplete recovery methods.
The article explains that Milvus 3.0 search aggregation operates on ANN candidates, not full collections, so missing buckets stem from candidate budget limits (topK and TopHits.size), not the safe_factor parameter; it distinguishes candidate-signal counts from exact filtered counts and advises choosing the right query path for each need.
This report validates a read-write separation architecture using Dameng DM8 for transactions and Easysearch for full-text search, syncing 1 million Chinese documents via Logstash; full sync completes in 5.5 minutes, incremental changes propagate within 30 seconds, and search latency drops 12x compared to SQL LIKE while maintaining identical recall.
This article details integrating Dameng DM8 database with Easysearch 2.4.0 via Logstash 9.0 for full-text search, covering version compatibility fixes, JDBC driver configuration, logical deletion handling, timestamp boundary issues, IK analyzer alignment for Chinese text, and benchmark results showing 17x faster queries with consistent recall versus LIKE.
This article details eight common SQL anti-patterns — including OFFSET pagination, implicit type conversion, correlated subqueries in updates, mixed sorting, EXISTS clauses, blocked condition pushdown, late filtering, and unoptimized intermediate results — with execution plans and rewrites that reduce query times from seconds to milliseconds.
In an 80-minute interview, Turing Award winner Mike Stonebraker argues that relational databases will absorb AI workloads, explains why Text-to-SQL fails on real enterprise data due to schema corruption and access controls, details DBOS's persistent workflow approach for AI agents, and discusses saga patterns for compensating transactions, graph database limitations, and the future of open-source AI.
Redis 8.0 introduces native vector search with Vector Set data type, HNSW indexing, quantization, and the Iris platform for agent memory and context retrieval, enabling semantic caching that cuts LLM costs by 70% and integrates with LangChain4j and Spring AI.
