Skip to content
plainsight.pro

Fabric Naming Conventions

Purpose

Standardized naming patterns create clarity and enable automation across the Fabric data estate. Through semantic prefixes, version indicators, and type codes, teams can automate governance processes and maintain clear data lineage. When implemented properly, these conventions accelerate development through discoverability while building trust through consistent, meaningful artifact names.

Overview

This section captures pragmatic rules you can enforce via templates and CI. Benefits and rationale are summarized below.

Benefit Why it matters
Discoverability Quickly locate related artifacts across workspaces and apps using predictable prefixes and purpose codes
Automation Scripts and CI can parse names to assign ownership, validate lifecycle, and enforce policies
Governance Regex-based policy checks make it possible to deny or flag non-conforming artifacts in PRs
Clarity Separates technical names from business-facing display names, reducing ambiguity

1datamarc_naming

Image source: Marc Lelijveld

Quick Reference: Do's and Don'ts

Do ✅ Don't ❌
Use semantic prefixes (e.g., DP_, LH_) for all artifacts Create artifacts without type prefixes
Keep purpose codes short and consistent (e.g., INGST, STORE) Use long, inconsistent purpose descriptions
Use underscores to separate prefix, purpose, and description Mix different separators (hyphens, dots) in names
Include the layer for lakehouse artifacts (e.g., _Bronze, _Silver) Skip the layer designation in lakehouse names
Use PascalCase or snake_case consistently within description Mix different naming cases in the same workspace
Keep names concise but meaningful Use generic names like "Test1" or overly long descriptions
Document new prefixes/purposes in a central location Create new prefixes without team alignment
Use automation to enforce naming standards Rely solely on manual reviews for naming compliance

Structure

Overall recommended structure:

{ItemType}_{Purpose}_{FreeTextDescription}

Position mapping:

1-2: Item type (2 chars) 3: underscore 4-8: Item purpose (up to ~5 chars recommended) 9: underscore 10+: free descriptive text (PascalCase or snake_case as you prefer)

This structure keeps type & purpose easily parseable for automation.

Item types (expanded list)

Prefix Item type Notes
CJ Copy Job Small ingest jobs
DP Data Pipeline Orchestration pipelines
DF Dataflow Dataflow assets
ES Eventstream Event ingestion definitions
MR Mirrored object External DB mirror
SD Spark Job Definition Spark job configs
NB Notebook Development notebooks
EN Environment Environment definitions
EX Experiment ML experiments
ML Machine Learning Model Trained models
LH Lakehouse OneLake/Lakehouse assets
WH Warehouse Fabric warehouse
EH Eventhouse Eventhouse items
DB SQL Database Synapse/SQL DBs
SM Semantic Model PBIP / model files
KQ KQL Queryset Kusto query sets
DA Data Agent Fabric data agent/skill
RP Report Power BI report
PR Paginated Report Paginated report
DS Dashboard Dashboard assets
RD Realtime Dashboard Real-time dashboards
SC Scorecard Scorecards
AC Activator Activation jobs
OA Org App Published apps
VL Variable Library Config/variables

Tip: choose a small subset of prefixes that match your org scale - you can expand as needed.

Item purposes

Purpose Meaning Example
ORCHS Orchestration chains DP_ORCHS_NightlyETL
INGST Ingestion CJ_INGST_Oracle
TRNSF Transform NB_TRNSF_CustomerCleansing
STORE Storage / persisting LH_STORE_Silver
ANLYZ Analytics SM_ANLYZ_YoYSales
SCIEN Data science / ML ML_SCIEN_ChurnModel
MAINT Maintenance tasks DP_MAINT_Vacuum
MONIT Monitoring DS_MONIT_Pipelines
CNFGS Configuration VL_CNFGS_Environments
DOCUM Documentation RP_DOCUM_DataCatalog

Examples

Artifact type Suggested name Why it helps
Pipeline (orchestration) DP_ORCHS_NightlyBatch Immediately shows it's an orchestration pipeline and nightly cadence
Copy job CJ_INGST_OracleDb Clear ingest job from Oracle DB
Lakehouse (silver) LH_STORE_Sales_Silver Shows domain (Sales) and medallion layer (Silver)
Semantic model SM_ANLYZ_SalesYTD Model focused on Sales year-to-date metrics
Report RP_ANLYZ_ExecSummary Business-facing Exec summary dashboard

Keep Environment Names Out of Artifact Names

Avoid embedding environment names into canonical artifact names (prefer Lakehouse_Bronze over Lakehouse_Bronze_DEV). Environment-aware behavior is better handled by workspace scoping, deployment pipelines, or display names. This keeps canonical names environment-agnostic and simplifies promotion across environments.

[!tip] Validating & Automating Naming Conventions - Validation: Use this regex pattern to validate names: ^[A-Z]{2}_[A-Z]{4,7}_[A-Za-z0-9_]+$ - Ownership: Assign ownership automatically using prefix + purpose (e.g., all LH_* items owned by the Data Platform team)