This is the authoring guide for the legacy dbt Semantic Layer YAML spec, supported in dbt Core 1.6 through 1.11. dbt Core 1.12+ also supports this spec but the latest spec is recommended for new projects.

In the legacy spec, semantic models are defined as top-level resources separate from dbt model definitions. Measures are a core concept used as building blocks for metrics.

Implementation Workflow (Steps 1-4: semantic model, entities, dimensions, measures/metrics)
YAML Format Reference (complete spec, measure properties, percentile, non-additive dimensions)
Metrics (simple, derived, cumulative, ratio, conversion)
SCD Type II Dimensions
Key Formatting Rules
Common Pitfalls

Implementation Workflow

Step 1: Define Semantic Model

Create a top-level semantic_models: entry that references the dbt model via ref(). Set defaults.agg_time_dimension to the primary time column. If the model does not have a time column, warn the user that it cannot contain time-based measures or metrics. Ask the user if they want to create a derived time dimension.

semantic_models:
  - name: orders
    model: ref('orders')
    defaults:
      agg_time_dimension: ordered_at

Step 2: Define Entities

Identify the primary key column. Add it to the entities: array with type: primary. Use expr to reference the actual column name when the entity name differs from it. If the model has foreign keys, define those as type: foreign.

semantic_models:
  - name: orders
    model: ref('orders')
    defaults:
      agg_time_dimension: ordered_at
    entities:
      - name: order
        type: primary
        expr: order_id
      - name: customer
        type: foreign
        expr: customer_id

Entity types: primary, foreign, unique, natural (SCD Type II only).

If the model has no physical primary key column, use the primary_entity property:

semantic_models:
  - name: bookings_monthly_source
    model: ref('bookings_monthly_source')
    defaults:
      agg_time_dimension: ds
    primary_entity: booking_id

Step 3: Define Dimensions

Scan columns for dimension candidates:

Time columns -> type: time with type_params.time_granularity
Categorical columns (strings, booleans) -> type: categorical

Use expr to reference the actual column name when the dimension name differs from it.

semantic_models:
  - name: orders
    model: ref('orders')
    defaults:
      agg_time_dimension: ordered_at
    entities:
      - name: order
        type: primary
        expr: order_id
      - name: customer
        type: foreign
        expr: customer_id
    dimensions:
      - name: ordered_at
        type: time
        type_params:
          time_granularity: day
      - name: order_status
        type: categorical

Computed dimensions use expr:

semantic_models: 
  - name: orders
    [...]
    dimensions:
      - name: is_bulk_transaction
        type: categorical
        expr: case when quantity > 10 then true else false end

Step 4: Define Measures and Metrics

Add a measures: array under the semantic model for aggregations. Then define top-level metrics: that reference those measures via type_params.

Supported agg types: sum, min, max, average, sum_boolean, count_distinct, median, percentile.

semantic_models:
  - name: orders
    model: ref('orders')
    defaults:
      agg_time_dimension: ordered_at
    entities:
      - name: order
        type: primary
        expr: order_id
      - name: customer
        type: foreign
        expr: customer_id
    dimensions:
      - name: ordered_at
        type: time
        type_params:
          time_granularity: day
      - name: order_status
        type: categorical
    measures:
      - name: order_count
        agg: sum
        expr: 1
      - name: total_revenue
        agg: sum
        expr: amount
      - name: average_order_value
        agg: average
        expr: amount

metrics:
  - name: order_count
    type: simple
    label: Order Count
    type_params:
      measure: order_count
  - name: total_revenue
    type: simple
    label: Total Revenue
    type_params:
      measure: total_revenue

YAML Format Reference

Complete Semantic Model Spec

semantic_models:
  - name: the_name_of_the_semantic_model   # Required
    description: same as always             # Optional
    model: ref('some_model')                # Required
    defaults:                               # Required
      agg_time_dimension: dimension_name    # Required if model contains measures
    entities:                               # Required
      - name: entity_name
        type: primary | foreign | unique | natural
        expr: column_name_or_sql_expression  # Optional, defaults to name
    dimensions:                             # Required
      - name: dimension_name
        type: categorical | time
        expr: column_name_or_sql_expression  # Optional, defaults to name
        type_params:                         # Required for time dimensions
          time_granularity: day
    measures:                               # Optional
      - name: measure_name
        agg: sum | min | max | average | sum_boolean | count_distinct | median | percentile
        expr: column_name_or_sql_expression  # Optional, defaults to name
    primary_entity: entity_name             # Required if no primary entity in entities array

Measure Properties

Property	Description	Required
`name`	Unique across all semantic models	Yes
`agg`	Aggregation type	Yes
`description`	Human-readable explanation	No
`expr`	Column name or SQL expression	No (defaults to name)
`label`	Display name in downstream tools	No
`create_metric`	Auto-generate a simple metric (`true`/`false`)	No
`agg_time_dimension`	Override default time dimension	No
`agg_params`	Extra params for specific agg types (e.g. percentile)	No
`non_additive_dimension`	For measures that shouldn't aggregate across time	No
`config`	Supports `meta` dictionary	No

Percentile Measures

semantic_models: 
  - name: orders
    [...]
    measures:
      - name: p99_transaction_value
        description: The 99th percentile transaction value
        expr: transaction_amount_usd
        agg: percentile
        agg_params:
          percentile: .99
          use_discrete_percentile: false

Non-Additive Dimensions

For measures like account balances or MRR that shouldn't be summed across time:

semantic_models: 
  - name: orders
    [...]
    measures:
      - name: mrr
        description: Sum of all active subscription plans
        expr: subscription_value
        agg: sum
        non_additive_dimension:
          name: subscription_date
          window_choice: max   # max (period end) or min (period start)
      - name: user_mrr
        description: Each user's MRR
        expr: subscription_value
        agg: sum
        non_additive_dimension:
          name: subscription_date
          window_choice: max
          window_groupings:
            - user_id

Metrics

All metrics are defined at the top-level metrics: key, referencing measures via type_params.

Simple Metrics

metrics:
  - name: customers
    description: Count of customers
    type: simple
    label: Count of customers
    type_params:
      measure:
        name: customers
        fill_nulls_with: 0
        join_to_timespine: true
        alias: customer_count
        filter: "{{ Dimension('customer__customer_total') }} >= 20"

Shorthand when no extra attributes needed:

metrics:
  - name: total_revenue
    type: simple
    label: Total Revenue
    type_params:
      measure: total_revenue

Derived Metrics

Combine multiple metrics using an expression.

metrics:
  - name: order_gross_profit
    description: Gross profit from each order.
    type: derived
    label: Order gross profit
    type_params:
      expr: revenue - cost
      metrics:
        - name: order_total
          alias: revenue
        - name: order_cost
          alias: cost

With offset window (period-over-period):

metrics:
  - name: order_total_growth_mom
    description: "Percentage growth of orders total compared to 1 month ago"
    type: derived
    label: Order total growth % M/M
    type_params:
      expr: (order_total - order_total_prev_month)*100/order_total_prev_month
      metrics:
        - name: order_total
        - name: order_total
          offset_window: 1 month
          alias: order_total_prev_month

With filter on input metric:

metrics:
  - name: food_order_gross_profit
    label: Food order gross profit
    type: derived
    type_params:
      expr: revenue - cost
      metrics:
        - name: order_total
          alias: revenue
          filter: |
            {{ Dimension('order__is_food_order') }} = True
        - name: order_cost
          alias: cost
          filter: |
            {{ Dimension('order__is_food_order') }} = True

Cumulative Metrics

Aggregate a measure over a running window or grain-to-date period. Requires a time spine.

metrics:
  - name: cumulative_order_total
    label: Cumulative order total (All-Time)
    description: The cumulative value of all orders
    type: cumulative
    type_params:
      measure:
        name: order_total

  - name: cumulative_order_total_l1m
    label: Cumulative order total (L1M)
    description: Trailing 1-month cumulative order total
    type: cumulative
    type_params:
      measure:
        name: order_total
      cumulative_type_params:
        window: 1 month

  - name: cumulative_order_total_mtd
    label: Cumulative order total (MTD)
    description: The month-to-date value of all orders
    type: cumulative
    type_params:
      measure:
        name: order_total
      cumulative_type_params:
        grain_to_date: month

With period_agg for re-aggregation at non-default granularity:

metrics:
  - name: cumulative_revenue
    description: The cumulative revenue for all orders.
    label: Cumulative revenue (all-time)
    type: cumulative
    type_params:
      measure: revenue
      cumulative_type_params:
        period_agg: first   # first | last | average. Defaults to first.

window and grain_to_date cannot be used together.

Ratio Metrics

Create a ratio between two metrics. Numerator and denominator can be strings (metric name) or dicts (with name, filter, alias).

metrics:
  - name: food_order_pct
    description: "The food order count as a ratio of the total order count"
    label: Food order ratio
    type: ratio
    type_params:
      numerator: food_orders
      denominator: orders

With filter and alias:

metrics:
  - name: frequent_purchaser_ratio
    description: Fraction of active users who qualify as frequent purchasers
    type: ratio
    type_params:
      numerator:
        name: distinct_purchasers
        filter: |
          {{Dimension('customer__is_frequent_purchaser')}}
        alias: frequent_purchasers
      denominator:
        name: distinct_purchasers

Conversion Metrics

Measure how often one event leads to another for a specific entity within a time window.

metrics:
  - name: visit_to_buy_conversion_rate_7d
    description: "Conversion rate from visiting to transaction in 7 days"
    type: conversion
    label: Visit to buy conversion rate (7-day window)
    type_params:
      conversion_type_params:
        base_measure:
          name: visits
          fill_nulls_with: 0
          filter: "{{ Dimension('visits__referrer_id') }} = 'facebook'"
        conversion_measure:
          name: buys
        entity: user
        window: 7 days

With constant properties:

metrics:
  - name: view_item_detail_to_purchase_with_same_item
    description: "Conversion rate for users who viewed and purchased the same item"
    type: conversion
    label: View item detail > Purchase
    type_params:
      conversion_type_params:
        calculation: conversions
        base_measure:
          name: view_item_detail
        conversion_measure: purchase
        entity: user
        window: 1 week
        constant_properties:
          - base_property: product
            conversion_property: product

SCD Type II Dimensions

For slowly changing dimension tables, use validity_params under type_params and natural entity type:

semantic_models:
  - name: sales_person_tiers
    description: SCD Type II table of tiers for salespeople
    model: ref('sales_person_tiers')
    defaults:
      agg_time_dimension: tier_start
    primary_entity: sales_person
    dimensions:
      - name: tier_start
        type: time
        label: "Start date of tier"
        expr: start_date
        type_params:
          time_granularity: day
          validity_params:
            is_start: True
      - name: tier_end
        type: time
        label: "End date of tier"
        expr: end_date
        type_params:
          time_granularity: day
          validity_params:
            is_end: True
      - name: tier
        type: categorical
    entities:
      - name: sales_person
        type: natural
        expr: sales_person_id

SCD Type II semantic models cannot contain measures.

Key Formatting Rules

Top-level semantic_models: key (not nested under models:)
model: ref('...') required on each semantic model without curly braces
defaults.agg_time_dimension for default time dimension
Entities, dimensions, and measures are separate arrays under the semantic model
All metrics at the top-level metrics: key, referencing measures via type_params
Use expr on dimensions/entities for computed values or column name aliasing

Common Pitfalls

Pitfall	Fix
Missing `defaults.agg_time_dimension`	Every semantic model with measures needs a default time dimension
`time_granularity` outside `type_params`	Must be nested under `type_params` for time dimensions
Missing `model: ref('...')`	Required for every semantic model
Metrics without `type_params`	All metrics must reference measures through `type_params`
Using `window` and `grain_to_date` together	Cumulative metrics can only have one
Missing `type_params.metrics` on derived metrics	Must list metrics used in `expr`
Using `semantic_model:` on models or `agg` on metrics	Those are latest spec syntax; this spec uses `semantic_models:` and `measures`

evals

skills

adding-dbt-unit-test

answering-natural-language-questions-with-dbt

building-dbt-semantic-layer

references

configuring-dbt-mcp-server

fetching-dbt-docs

migrating-dbt-core-to-fusion

running-dbt-commands

troubleshooting-dbt-job-errors

using-dbt-for-analytics-engineering

CONTRIBUTING.md

README.md

tile.json

dbt-labs/dbt-agent-skills

legacy-spec.mdskills/building-dbt-semantic-layer/references/

Legacy Semantic Layer YAML Spec

Contents

Implementation Workflow

Step 1: Define Semantic Model

Step 2: Define Entities

Step 3: Define Dimensions

Step 4: Define Measures and Metrics

YAML Format Reference

Complete Semantic Model Spec

Measure Properties

Percentile Measures

Non-Additive Dimensions

Metrics

Simple Metrics

Derived Metrics

Cumulative Metrics

Ratio Metrics

Conversion Metrics

SCD Type II Dimensions

Key Formatting Rules

Common Pitfalls

dbt-labs/dbt-agent-skills

legacy-spec.md.css-3qkkll{font-size:var(--chakra-font-sizes-sm);font-weight:var(--chakra-font-weights-normal);color:var(--chakra-colors-gray-300);}skills/building-dbt-semantic-layer/references/

Legacy Semantic Layer YAML Spec

Contents

Implementation Workflow

Step 1: Define Semantic Model

Step 2: Define Entities

Step 3: Define Dimensions

Step 4: Define Measures and Metrics

YAML Format Reference

Complete Semantic Model Spec

Measure Properties

Percentile Measures

Non-Additive Dimensions

Metrics

Simple Metrics

Derived Metrics

Cumulative Metrics

Ratio Metrics

Conversion Metrics

SCD Type II Dimensions

Key Formatting Rules

Common Pitfalls

legacy-spec.mdskills/building-dbt-semantic-layer/references/