ACID
Atomicity, Consistency, Isolation, Durability - the four correctness guarantees of a relational transaction. Used as shorthand for "transactionally safe" storage, as opposed to BASE-style eventual consistency.
BASE
Basically Available, Soft state, Eventual consistency - the relaxed consistency model favoured by many distributed NoSQL systems. It contrasts with ACID by trading immediate consistency for availability and partition tolerance.
BCNF
Boyce-Codd Normal Form - a stricter version of third normal form requiring that every non-trivial functional dependency have a superkey on its left-hand side. A schema in BCNF eliminates redundancy from all candidate-key anomalies.
CTE
Common Table Expression - a named temporary result set introduced with WITH, scoped to a single statement, used to structure complex queries and to express recursion via WITH RECURSIVE.
DBA
Database Administrator - the role responsible for installing, configuring, securing, backing up, tuning, and monitoring database systems. DBAs own capacity planning, replication topology, and disaster recovery.
DCL
Data Control Language - the SQL subset that manages permissions, principally GRANT and REVOKE. It governs who can perform which operations on which objects.
DDL
Data Definition Language - the SQL subset that defines and alters schema objects: CREATE, ALTER, DROP, TRUNCATE. It changes structure rather than data.
DML
Data Manipulation Language - the SQL subset that queries and modifies row data: SELECT, INSERT, UPDATE, DELETE, MERGE. These statements run inside transactions.
DQL
Data Query Language - sometimes carved out from DML to name the read-only SELECT statement specifically, the declarative core of SQL.
ER / ERD
Entity-Relationship / Entity-Relationship Diagram - a conceptual model and its diagram, depicting entities, their attributes, and the relationships and cardinalities among them. ERDs are the standard artefact for designing a relational schema.
ETL
Extract, Transform, Load - the pipeline pattern that pulls data from sources, reshapes it, and writes it into a target store such as a data warehouse. ELT is the variant that loads first and transforms inside the warehouse.
FK
Foreign Key - a column referencing another table's primary or unique key, enforcing referential integrity between related tables.
GIN
Generalized Inverted Index - a PostgreSQL index type optimised for values containing multiple component keys, such as full-text documents, arrays, and jsonb. It powers @@ text search and containment operators.
HNSW
Hierarchical Navigable Small World - a graph-based approximate nearest neighbour index used for vector similarity search, available via the pgvector extension and in dedicated vector databases. It enables fast top-k semantic retrieval.
JDBC
Java Database Connectivity - the standard Java API for connecting to relational databases through driver implementations, exposing connections, statements, and result sets. It underpins JPA/Hibernate and most JVM data access.
MVCC
Multi-Version Concurrency Control - a concurrency mechanism that keeps multiple row versions so readers see a consistent snapshot without blocking writers. Used by PostgreSQL and Oracle.
OLAP
Online Analytical Processing - analytical, scan-and-aggregate workloads over large historical datasets, typically served by columnar or warehouse engines.
OLTP
Online Transaction Processing - high-volume, low-latency transactional workloads of many small concurrent reads and writes, typically served by row-oriented relational databases.
ORM
Object-Relational Mapping - a layer that maps database tables to in-language objects and generates SQL, e.g. Hibernate, Prisma, SQLAlchemy, Entity Framework.
PK
Primary Key - the column(s) that uniquely identify each row, enforced as NOT NULL and unique; the target of foreign-key references.
RDBMS
Relational Database Management System - software that stores data in relations (tables) and implements SQL and ACID transactions, e.g. PostgreSQL, MySQL, SQL Server, Oracle.
TCL
Transaction Control Language - the SQL subset that manages transactions: BEGIN/START TRANSACTION, COMMIT, ROLLBACK, SAVEPOINT.
TPS / QPS
Transactions Per Second / Queries Per Second - throughput metrics for transactional and query load respectively, used in capacity planning and benchmarking.
UDF
User-Defined Function - a custom function written in SQL or a procedural language (PL/pgSQL, PL/SQL, T-SQL) and callable in queries, extending the built-in function set.
UPSERT
Insert-or-update - a single operation that inserts a row, or updates it if a key conflict occurs, expressed as INSERT ... ON CONFLICT DO UPDATE (PostgreSQL) or MERGE / INSERT ... ON DUPLICATE KEY UPDATE (others).
UUID
Universally Unique Identifier - a 128-bit identifier (often v4 random or v7 time-ordered) used as a globally unique surrogate key, avoiding coordination across systems at the cost of size and index locality.
WAL
Write-Ahead Log - the append-only log of changes written before data files are updated, providing crash recovery, point-in-time recovery, and the stream for physical replication. Called the redo log or transaction log in other engines.
1NF / 2NF / 3NF
First, Second, and Third Normal Form - successive normalization stages: 1NF requires atomic column values, 2NF removes partial dependencies on part of a composite key, 3NF removes transitive dependencies on non-key columns.