GBDS flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Business requirement

A statement of a high-level organizational need or objective that justifies a project. It answers why the work is being done, independent of any particular solution.

Non-functional requirement

A quality or constraint the solution must satisfy, such as performance, availability, security, or usability. It describes how well the system behaves rather than what function it performs.

Traceability matrix

A grid linking each requirement to its source, design components, and test cases. It exposes requirements with no test coverage and tests with no originating requirement.

MoSCoW prioritization

A ranking scheme that classifies scope as Must have, Should have, Could have, or Won't have this time. If any Must item is dropped, the delivery is considered a failure.

User story

A short, user-centered statement of a need, typically formatted as role, goal, and benefit. It is a placeholder for conversation, refined by acceptance criteria before development.

Third normal form

A table is in 3NF when it is in 2NF and no non-key attribute depends transitively on the primary key. This removes redundancy where a non-key attribute determines another.

Associative entity

A table created to resolve a many-to-many relationship. It holds foreign keys to both parent entities and may carry attributes belonging to the relationship itself.

Surrogate key

A system-generated identifier with no business meaning, such as an auto-incremented integer. It insulates the model from changes to natural business keys.

Star schema

A dimensional design with a central fact table holding measures and foreign keys, surrounded by denormalized dimension tables. It favors simple, fast analytical queries over normalization.

Cardinality

The maximum number of instances of one entity that can relate to an instance of another, expressed as one-to-one, one-to-many, or many-to-many. It is distinct from optionality.

Slowly changing dimension Type 2

A dimension technique that preserves history by inserting a new row for each change, with effective start and end dates and a current-record flag. Each version keeps its own surrogate key.

Primary key

A column or set of columns that uniquely identifies each row in a table and cannot be null. A table has at most one primary key, though it may span multiple columns.

HAVING clause

A SQL clause that filters grouped rows after aggregation, so it can reference aggregate functions. It is applied after GROUP BY, unlike WHERE, which filters individual rows first.

LEFT OUTER JOIN

A join returning all rows from the left table plus matching rows from the right, with nulls where no match exists. A WHERE filter on right-table columns can silently remove those null rows.

Transaction isolation level

A setting controlling how much concurrent transactions can see each other's uncommitted or changing data. Higher levels prevent more anomalies but reduce concurrency and can cause blocking.

NULL semantics

NULL represents an unknown or absent value, so equality comparisons with it return UNKNOWN rather than true. Use IS NULL to test for it, and remember aggregates skip nulls.

Clustered index

An index that determines the physical storage order of table rows, so a table can have only one. Lookups by the clustering key are efficient because rows are stored together.

Data quality dimension

A measurable aspect of data fitness such as completeness, accuracy, consistency, timeliness, or uniqueness. Each dimension needs a concrete rule and threshold to be monitored objectively.

Data steward

A role responsible for the day-to-day quality, definitions, and correct use of data within an assigned domain. Stewards execute governance decisions made by data owners.

Data lineage

Documentation of the origin and movement of data from source systems through transformations to reports. It supports impact analysis, root-cause investigation, and regulatory audit.

Data masking

Replacing sensitive values with realistic but altered substitutes so non-production users see no real personal data. Static masking alters stored copies; dynamic masking applies at query time.

Principle of least privilege

Granting each user or process only the access required to perform its function, and no more. It limits the blast radius of compromised accounts and accidental errors.

Change data capture

A technique that identifies and extracts only rows inserted, updated, or deleted since the last load, often by reading transaction logs. It captures hard deletes that timestamp comparisons miss.

Idempotent load

A load process that produces the same target state whether run once or many times, typically via merge or partition overwrite. It makes reruns after failure safe.

Inferred member

A placeholder dimension row created when a fact arrives before its dimension record. It lets the fact load proceed and is later updated when the real dimension data appears.

Data profiling

Statistical analysis of source data to reveal null rates, distinct counts, value ranges, and pattern distributions. Results inform transformation rules and quality thresholds before development.

Semi-additive measure

A fact-table measure that can be summed across some dimensions but not all, typically not across time. Account balances are the classic example and require snapshot logic.

Conformed dimension

A dimension table reused across multiple fact tables with the same keys, attributes, and definitions. It enables consistent drill-across reporting and comparable metrics.

Fact table grain

The precise meaning of one row in a fact table, such as one order line per day. Declaring grain first prevents mixed levels of detail and double counting.

Key performance indicator

A metric tied to a strategic objective with a defined formula, target, and owner. Without a stated denominator and threshold, a metric cannot drive a decision.

Critical path

The longest chain of dependent tasks through a project network, which determines the shortest possible duration. Any delay on this path delays the whole project.

Cost performance index

Earned value divided by actual cost, showing cost efficiency on the project. A value below 1.0 means work is costing more than planned for the value delivered.