Vulcan Terminology Guide¶

When you're working with Vulcan, you'll encounter terms that might be unfamiliar. This guide explains what they mean and why they matter. Each term links to detailed documentation where you can dive deeper.

Model Terms¶

These control how your models behave and how Vulcan processes them.

Term	Definition	Documentation
Grain	The column or columns that make each row unique. Like declaring a primary key, but Vulcan uses this for table comparisons and semantic layer joins.	Model Properties
References	Foreign key columns that tell Vulcan how models relate to each other. The semantic layer uses these to detect joins automatically.	Model Properties
Model Kind	How Vulcan processes your data. FULL rebuilds everything, INCREMENTAL only processes new intervals, VIEW computes on-demand.	Model Kinds
Cron	Schedule expression that sets when your model runs. Use `@daily` for nightly refreshes or standard cron syntax for custom schedules.	Model Properties
Interval	A time period Vulcan tracks for incremental models. Each day, week, or hour becomes an interval that gets processed independently.	Incremental Models
Backfill	Loading historical data when you first create a model or after changing logic. Vulcan figures out which intervals need processing.	Plans
Lookback	How many intervals to reprocess for late-arriving data. Handles events that show up after their time window already passed.	Model Kinds
Forward-only	Models that never rebuild historical data. Use when past data is immutable or reprocessing costs too much.	Model Kinds
Owner	Team or person responsible for a model. Used for notifications and knowing who to contact when something breaks.	Model Properties
Description	Human-readable explanation of what your model does. Vulcan registers this as a table comment, so it shows up in BI tools.	Model Properties
Depends On	Explicit dependency list. Required for Python models since Vulcan can't auto-detect dependencies from Python code.	Model Properties

Execution Terms¶

These describe how Vulcan applies changes and processes data.

Term	Definition	Documentation
Plan	Summary of what will change before you deploy. Shows affected models, intervals that need processing, and the full impact of your changes.	Plans
Run	Scheduled execution that processes new data intervals. Different from plan: plan applies code changes, run handles regular data processing.	Run and Scheduling
Virtual Environment	Isolated testing space that can share tables when safe. Test changes without touching production, roll back if needed.	Plans
Virtual Layer	Views that point to physical tables. This is what you interact with most of the time, not the raw storage underneath.	Plans
Physical Layer	Actual database tables and materialized views where your data lives. Vulcan manages this automatically.	Plans

Semantic Layer Terms¶

These transform your technical tables into something business users can query without SQL.

Term	Definition	Documentation
Semantic Layer	Translation layer that maps physical models to business concepts. Turns `analytics.daily_revenue_metrics` into "Monthly Revenue by Customer Tier" that anyone can query.	Semantic Layer Overview
Semantic Model	Maps a physical Vulcan model to a semantic representation. Provides business aliases and exposes dimensions, measures, segments, and joins.	Semantic Models
Dimensions	Attributes for grouping and filtering. Answer "by what?" questions. Your model columns become dimensions automatically.	Semantic Models
Measures	Aggregated calculations. Answer "how much?" or "how many?" using SQL expressions like `SUM(amount)` or `COUNT(*)`.	Semantic Models
Segments	Reusable filter conditions. Define subsets like "active customers" or "high-value orders" that you can reuse across queries.	Semantic Models
Joins	Relationships between semantic models. Enable cross-model analysis, like combining order data with customer data automatically.	Semantic Models
Business Metrics	Complete analytical definitions that combine measures with dimensions and time. Ready for time-series analysis and automatic API generation.	Business Metrics
Transpilation	Convert semantic queries to executable SQL. Validate and debug business-friendly queries before they hit your database.	Transpiling Semantics

Data Quality Terms¶

These ensure data integrity. The difference between catching problems early and discovering them at 3 AM.

Term	Definition	Documentation
Assertions / Audits	SQL queries that validate data after execution. Always blocking: if they find bad data, execution stops immediately.	Audits
Checks	Quality monitoring rules configured in YAML. Non-blocking validation that tracks trends and detects anomalies over time.	Data Quality

Architecture Terms¶

These describe how Vulcan structures and tracks your pipeline.

Term	Definition	Documentation
DAG	Directed Acyclic Graph. Structure Vulcan uses to track model dependencies and figure out the correct execution order.	DAG
Lineage	Visualization of how data flows from sources to consumption. See how changes propagate through your pipeline.	Lineage

How These Terms Fit Together¶

You don't need to memorize everything. Here's how these concepts connect in practice:

→ Models define your logic

→ Grain documents structure

→ Model Kind determines processing

→ Cron sets schedule

→ Plan reviews changes

→ Run processes data

→ Assertions validate quality

→ Semantic Layer exposes business interface

→ Dimensions and Measures enable self-service analytics

Each term solves a specific problem. Grain helps with comparisons. References enable automatic joins. Model Kinds optimize performance. The Semantic Layer makes data accessible to non-technical users.

Click any documentation link above for detailed guides with examples, best practices, and advanced patterns.