Overview

This document outlines the standardized naming conventions for Microsoft Fabric resources across our data engineering ecosystem. Following these conventions ensures consistency, clarity, and ease of navigation across our Medallion Architecture implementation and Deployment process.

Workspace Naming

Workspaces serve as the primary containers for our data assets. We maintain three distinct workspace types:

Workspace Type	Naming Pattern	Example	Purpose
Source	src_[source_name]_[deployment stage]	src_dalux_dev	Contains raw ingested data from a single source. Only data available from the source
Project	prj_[project name]_[deployment stage]	prj_mor_dev	Houses project-specific data transformations
BI	bi_[application name]_[deployment stage]	bi_project_delivery_dev	Dedicated to semantic models and semantic models

Lakehouse Naming

Lakehouses represent our primary storage and processing layer within each workspace:

Layer	Naming Pattern	Location	Example	Purpose
Bronze and Silver	src_lh_[source_name]	Source Workspace	src_lh_bimcollab	Raw data from source systems
Gold	gold_lh_[project_name]	Project Workspace	gold_lh_permits	Business-ready, project specific logics, project specific data

Schema Naming

Schemas help organize tables within lakehouses based on their processing stage and data domain. Our schema strategy reflects both the medallion architecture layers and the source origins of data.

Schema Type	Naming Pattern	Location	Example	Purpose
Bronze	bronze	Source Lakehouse	bronze	Contains raw, unmodified data from source systems
Silver	silver	Source Lakehouse	silver	Contains cleaned, validated, and standardized data
Source Shortcut	src_[source_name]	Project Lakehouse	src_dalux	Contains shortcuts to silver tables from specific sources
Gold	gold	Project Lakehouse	gold	Contains project-specific business logic and transformations
Utility	util	Any Lakehouse	util	Contains utility tables, metadata, and audit information

The schema structure helps maintain clear boundaries between medallion layers while providing context on data origins. For project lakehouses, the source-specific shortcut schemas (src_[source_name]) make it easy to identify which source a particular table originated from, especially important when a project combines data from multiple sources.

Table Naming

Tables follow consistent patterns based on their processing stage and business purpose:

Layer	Table Type	Naming Pattern	Example	Description
Bronze	Source	raw_[entity]	raw_projects	Unmodified source data
Silver	Cleansed	cl_[entity]	cl_projects	Cleaned, typed, and validated data
Silver	Conformed	cf_[entity]	cf_projects	Standardized business entities
Gold	Dimension	dim_[dimension]	dim_project	Dimensional model dimension tables
Gold	Fact	fact_[business process]	fact_project_changes	Dimensional model fact tables
Gold	Aggregate	agg_[measure]_[grain]	agg_project_status_monthly	Pre-aggregated metrics
Any	Reference	ref_[entity]	ref_status_codes	Reference or lookup data
Any	Temporary	tmp_[entity]_[purpose]	tmp_projects_dedupe	Temporary processing tables

Shortcut tables should maintain the same name as their source tables to preserve clarity and context.

Column Naming

Key Columns

Column Type	Naming Pattern	Example	Purpose
Primary Key	[entity]_id	project_id	Natural business key
Surrogate Key	[entity]_key	project_key	Generated surrogate key
Foreign Key	[referenced entity]_id	customer_id	Reference to another entity
Foreign Surrogate Key	[referenced entity]_key	customer_key	Reference to surrogate key

Common Data Elements

Column Type	Naming Pattern	Examples
Timestamps	[event]_timestamp	created_timestamp, modified_timestamp
Dates	[event]_date	start_date, due_date, birth_date
Flags	is_[state]	is_active, is_deleted, is_primary
Amounts	[type]_amount	invoice_amount, tax_amount
Counts	[item]_count	record_count, error_count
Status	[entity]_status	project_status, order_status

Technical Columns

Column Type	Naming Pattern	Examples
Partition Columns	partition_[type]	partition_date, partition_region
Audit Columns	etl_[event]_[detail]	etl_insert_timestamp, etl_source_system
Version Control	record_version	Records the version number of the record
Deleted Flag	is_deleted	Soft delete indicator
Validity Period	valid_from_date, valid_to_date	For SCD Type 2 dimensions

JSON/Complex Types

Column Type	Naming Pattern	Examples
Raw JSON	[entity]_json	event_json, payload_json
Struct Types	[domain]_details	address_details, contact_details
Array Types	[item]_array	product_array, tag_array

This is a living document. Update your naming conventions as your data platform evolves.

Naming Conventions