Persistent Data Management — Glossary

Term	Meaning
ACID	Atomicity, consistency, isolation, and durability properties of transactions
Active domain	Values that actually occur in a database instance or query constants
Attribute	Named property/column of an entity or relation
Attribute closure	Attributes implied by an attribute set under functional dependencies
B+ tree	Balanced index tree with all data entries in linked leaves
BCNF	Normal form in which every non-trivial FD determinant is a super key
Candidate key	Minimal set of attributes uniquely identifying a tuple
Cardinality	Relationship mapping ratio or, in query plans, estimated/result row count
Cascadeless schedule	Schedule in which transactions read only committed values
Checkpoint	Recovery marker that limits log scanning
Clustered index	Index whose order determines or closely matches row storage order
Commit	Successful durable completion of a transaction
Conflict	Same item, different transactions, with at least one write
Conflict serializability	Equivalence to a serial schedule by swapping non-conflicting operations
Cursor	Mechanism for processing query results one row at a time
Data independence	Ability to change one schema level with little/no impact above it
DDL	SQL commands that define schema objects
Deadlock	Cycle of transactions each waiting for another
Dense index	Index with an entry for every search-key value/record
Domain	Permitted set of values for an attribute
DRC	Relational calculus whose variables range over domain values
Durability	Committed transaction effects survive failure
Entity	Distinguishable real-world object
Entity integrity	Primary-key values cannot be null
Execution plan	Physical operators chosen to run a query
Foreign key	Attribute set referencing a candidate key in another/same relation
Fragmentation	Splitting a distributed relation horizontally, vertically, or both
Functional dependency	Constraint X→Y in which X values determine Y values
Hashing	Mapping a key to a bucket using a hash function
HDFS	Distributed, replicated block-oriented Hadoop file system
Index	Auxiliary access structure that speeds data retrieval
Isolation	Degree to which concurrent transactions are separated
Join	Operation combining related tuples from relations
Lock	Concurrency-control claim on a data resource
Lossless join	Decomposition whose join reconstructs exactly the original relation
Materialized view	Physically stored result of a query
MVD	Multivalued dependency expressing independent sets of values
MVCC	Concurrency method maintaining multiple data versions
Normalization	Decomposing relations to reduce redundancy and anomalies
Optimizer	DBMS component that selects an estimated low-cost plan
Phantom	New/deleted qualifying row observed when a predicate is repeated
Precedence graph	Directed conflict graph used to test serializability
Primary key	Candidate key chosen as the main tuple identifier
Projection	Relational operation selecting attributes/columns
Query selectivity	Fraction of input rows retained by a predicate
Recoverable schedule	Dependent transaction commits only after the source writer commits
Referential integrity	Foreign key matches a referenced value or permitted null
Relation	Set of tuples over named attributes; represented as a table
Relational algebra	Procedural formal language of relation operations
Relational calculus	Declarative formal query language based on logic
Replication	Maintaining copies of data at multiple sites
Rollback	Undoing transaction changes
Safe expression	Calculus query guaranteed a finite, domain-restricted result
Schedule	Interleaving of transaction operations preserving each transaction's order
Selection	Relational operation choosing rows that satisfy a predicate
Serializability	Correctness criterion requiring effect equivalent to a serial order
Sparse index	Ordered index containing entries for only selected keys/pages
Spark	Distributed DAG-based data-processing engine
Spurious tuple	Incorrect tuple introduced by joining a lossy decomposition
Strict schedule	Uncommitted writes cannot be read or overwritten by another transaction
Super key	Any attribute set that uniquely identifies a tuple
Timestamp ordering	Concurrency protocol enforcing a timestamp-based serial order
Transaction	Logical unit of database work
Trigger	Program automatically fired by a database event
TRC	Relational calculus whose variables range over tuples
Two-Phase Commit	Distributed atomic-commit protocol with prepare and decision phases
Two-Phase Locking	Protocol with a lock-growing phase followed by a shrinking phase
View	Named stored query presenting a virtual table
WAL	Write-ahead logging: log reaches stable storage before corresponding data
Weak entity	Entity identified using an owner key plus a partial key