Informatica Data Engineering Foundation key terms

A reusable design object in PowerCenter that defines how source data is read, transformed, and written to a target. It contains transformations connected by pipelines and is the basic unit of data integration logic.

A PowerCenter object that changes, routes, joins, aggregates, or otherwise processes data as it moves through a mapping. Each transformation has ports for input, output, and sometimes both, plus configurable properties.

A task that executes a single mapping against a configured source and target connection, with runtime options such as commit intervals, error thresholds, and partitioning. Sessions run inside workflows.

A container that sequences tasks such as sessions, command tasks, and decision tasks, with links and conditions controlling order. It is scheduled or started manually by the Integration Service.

The PowerCenter component that runs workflows and sessions, connects to sources and targets, and applies transformations. It can run on multiple nodes and supports load balancing and failover.

The PowerCenter component that manages the metadata repository database, controlling access, locking, and consistency for folders, mappings, and other objects. Clients connect through it rather than to the database directly.

A passive transformation that calculates new values row by row using expressions, such as concatenations, date math, or conditional logic. It can output both calculated and passed-through ports.

An active transformation that groups rows and computes aggregates such as sum, count, average, minimum, and maximum. Group-by ports define the groups, and sorted input can improve performance.

A transformation that retrieves values from a lookup table, file, or pipeline and returns them to the current row. It supports connected or unconnected use and can cache data for performance.

A transformation that joins two pipelines within a mapping, supporting inner, outer, and full joins. Unlike a database join, it requires a master and detail source and is often used across heterogeneous systems.

An active transformation that evaluates rows against multiple groups and routes each row to the first matching group or to a default group. It replaces chains of filters for conditional splitting.

Values defined at the mapping level that are supplied at runtime (parameters) or changed during a session (variables). They let one mapping serve multiple environments or date ranges without editing logic.

A technique that converts mapping logic into SQL and runs it in the source or target database, reducing data movement. It is most effective when transformations can be expressed in the database dialect.

A method of identifying and extracting only rows that changed since the last run, using timestamps, sequence numbers, or database logs. It reduces load volume and supports incremental integration.

An integration pattern that processes only new or changed records rather than a full extract. It typically relies on change data capture logic, high-water marks, or comparison with a prior snapshot.

A defined check that data must satisfy, such as a non-null constraint, a value range, a pattern match, or a referential relationship. Rules are scored and monitored to measure conformance.

The process of analyzing data to discover its structure, completeness, uniqueness, value distributions, and relationships. Profiling results inform rule design and reveal anomalies before they reach consumers.

A summarized view of data quality results across rules, dimensions, or objects, often expressed as scores or pass rates over time. It helps stakeholders track trends and prioritize remediation.

The workflow for recording, assigning, investigating, and resolving records that fail a data quality rule. It closes the loop between detection and correction and provides an audit trail.

The practice of continuously monitoring data pipelines and datasets for freshness, volume, schema, distribution, and quality anomalies. It aims to detect and explain issues before downstream users are affected.

A record of where data comes from, how it is transformed, and where it flows, at column or dataset level. It supports impact analysis, root-cause investigation, and regulatory documentation.

A controlled set of codes and descriptions used across systems, such as country codes, currency codes, or status values. Consistent reference data prevents mismatches in joins and reporting.

The core business entities shared across systems, such as customers, products, suppliers, and locations. It is relatively stable, authoritative, and reused by many processes and applications.

The single trusted version of a master data entity, assembled from contributing sources and resolved through matching and survivorship rules. It is published to consumers as the authoritative view.

The MDM process that identifies records referring to the same real-world entity and combines them. Matching uses keys and fuzzy rules; merging applies survivorship to build a consolidated record.

The set of rules that decides which source value wins for each attribute when records are consolidated. Rules may favor the most recent, most trusted, or most complete source.

A numeric indicator of how reliable a source system or record is, used in survivorship and matching decisions. Higher-trust sources are more likely to contribute values to the golden record.

A parent-child relationship among master data entities, such as an organizational structure or product category tree. MDM systems model and maintain these relationships alongside entity attributes.

A modeled object in an MDM hub that represents a real-world concept, such as Customer or Product, with its attributes, relationships, and validation rules. It is the base for master data management.

A role accountable for the quality, definition, and appropriate use of a data domain. Stewards resolve quality issues, approve changes, and connect business meaning with technical implementation.

The role accountable for a data domain's policy decisions, access approvals, and overall fitness for use. Owners set direction, while stewards handle day-to-day quality and definition work.

The framework of roles, policies, processes, and metrics that directs how data is defined, protected, and used. It balances business value with risk, compliance, and accountability.

The practice of labeling data by sensitivity and business criticality, such as public, internal, confidential, or restricted. Labels drive access controls, retention, and handling requirements.

A technique that replaces sensitive values with realistic but altered data, either statically in a copy or dynamically at query time. It protects privacy while preserving usability for testing and analytics.

A curated list of business terms with approved definitions, synonyms, owners, and related policies. It aligns stakeholders on meaning and links concepts to technical assets and quality rules.

Descriptive information about data assets and processes, such as schemas, data types, mappings, jobs, and lineage. It enables discovery, impact analysis, and automation across the data estate.

A searchable inventory of data assets that combines technical metadata, business descriptions, ownership, and usage information. It helps users find, understand, and trust data before using it.

Any managed item that holds or describes data, such as a table, file, report, dashboard, mapping, or API. Cataloging assets with owners and descriptions makes them discoverable and governable.