code (4)

These rules inspect the raw SQL/Jinja code of your dbt resources.

Rule: `code_contains_refs`

Manifest Rule

code_contains_refs details

This rule checks that the raw SQL code of a resource contains at least one ref() or source() function call. It inspects the actual code text (not the dependency graph), stripping SQL comments before checking. Detection is case-insensitive.

Resources without these calls may be hardcoding warehouse table names and bypassing dbt lineage tracking. Also check out has_refs, which checks the dependency graph (depends_on.nodes) rather than the raw SQL text. It differs since has_refs also checks the depends_on config.

Configuration

type: Must be code_contains_refs.
applies_to: (optional) List of dbt object types to include.
- Default: ["models", "snapshots"]
- Options: models, snapshots, macros

Common Rule Config

name: Human-readable name of the rule.
severity: "error" (fail) or "warning" (warn only).
- (optional, defaults to "error" if not specified)
description: Human-readable explanation of the rule.
category: Override the default rule category. Included in structured output (JSON, CSV, NDJSON) but not in the CLI table. Each rule has a built-in default (e.g. documentation, naming, testing, governance, structure, performance).
- (optional, defaults to the rule type's built-in category)
includes: List of patterns to explicitly include for this rule. See Includes & Excludes for pattern syntax and examples.
excludes: List of patterns to explicitly exclude from this rule. See Includes & Excludes for pattern syntax and examples.
model_materializations: Filter models by materialization type. Only applies when applies_to includes models.
- (optional, if not specified all materializations are included)
- Built-in types: table, view, incremental, ephemeral, materialized_view. Custom materializations are also supported.
- Example: ["table", "incremental"]

Example Config

manifest_tests:
  - name: "code_must_use_refs"
    type: "code_contains_refs"
    description: "All SQL code must contain ref() or source() calls."
    # severity: "warning"  (optional)
    # applies_to: ['models', 'snapshots']  (optional)
    # includes: ["models/staging/*"]
    # excludes: ["models/base/*"]

[[manifest_tests]]
name = "code_must_use_refs"
type = "code_contains_refs"
description = "All SQL code must contain ref() or source() calls."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["models/staging/*"]
# excludes = ["models/base/*"]

[[tool.dbtective.manifest_tests]]
name = "code_must_use_refs"
type = "code_contains_refs"
description = "All SQL code must contain ref() or source() calls."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["models/staging/*"]
# excludes = ["models/base/*"]

Relevant dbt code

-- models/good_model.sql (PASS - contains ref())
SELECT
    customer_id,
    first_name
FROM {{ ref('stg_customers') }}

-- models/good_source_model.sql (PASS - contains source())
SELECT *
FROM {{ source('raw', 'customers') }}

-- models/bad_model.sql (FAIL - hardcoded table name)
SELECT id, name
FROM raw_schema.users
WHERE active = true

-- models/tricky_model.sql (FAIL - ref is commented out)
-- SELECT * FROM {{ ref('stg_users') }}
SELECT id FROM raw_schema.users

Rule: `has_forbidden_code`

Manifest Rule

has_forbidden_code details

This rule checks if code contains forbidden patterns. Use it to enforce coding standards by flagging undesired patterns such as `SELECT *` statements, hardcoded references, or any other string patterns that should not appear in your dbt code.

Configuration

type: Must be has_forbidden_code.
forbidden_patterns: List of string patterns that are not allowed in the code. Each pattern is matched as a substring.
case_sensitive: (optional) Whether pattern matching should be case-sensitive. Defaults to false (case-insensitive).
applies_to: (optional) List of dbt object types to check.
- Default: ["models", "snapshots", "macros"]
- Options: models, snapshots, macros

Common Rule Config

name: Human-readable name of the rule.
severity: "error" (fail) or "warning" (warn only).
- (optional, defaults to "error" if not specified)
description: Human-readable explanation of the rule.
category: Override the default rule category. Included in structured output (JSON, CSV, NDJSON) but not in the CLI table. Each rule has a built-in default (e.g. documentation, naming, testing, governance, structure, performance).
- (optional, defaults to the rule type's built-in category)
includes: List of patterns to explicitly include for this rule. See Includes & Excludes for pattern syntax and examples.
excludes: List of patterns to explicitly exclude from this rule. See Includes & Excludes for pattern syntax and examples.
model_materializations: Filter models by materialization type. Only applies when applies_to includes models.
- (optional, if not specified all materializations are included)
- Built-in types: table, view, incremental, ephemeral, materialized_view. Custom materializations are also supported.
- Example: ["table", "incremental"]

Example Config

manifest_tests:
  # Forbid SELECT * in models (case-insensitive by default)
  - name: "no_select_star"
    type: "has_forbidden_code"
    forbidden_patterns: ["SELECT *"]
    description: "Models should not use SELECT *."
    # case_sensitive: false  (optional, default)
    # severity: "warning"  (optional)
    # applies_to: ['models', 'snapshots'] (optional)
    # includes: ["path/to/include/*"]
    # excludes: ["path/to/exclude/*"]

  # Case-sensitive match for exact patterns
  - name: "no_hardcoded_schema"
    type: "has_forbidden_code"
    forbidden_patterns: ["raw_prod.", "analytics_prod."]
    case_sensitive: true
    severity: "warning"
    description: "Use dbt selectors"

# Forbid SELECT * in models (case-insensitive by default)
[[manifest_tests]]
name = "no_select_star"
type = "has_forbidden_code"
forbidden_patterns = ["SELECT *"]
description = "Models should not use SELECT *."
# case_sensitive = false  # (optional, default)
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

# Case-sensitive match for exact patterns
[[manifest_tests]]
name = "no_hardcoded_schema"
type = "has_forbidden_code"
forbidden_patterns = ["raw_prod.", "analytics_prod."]
case_sensitive = true
severity = "warning"
description = "Use dbt selectors"

# Forbid SELECT * in models (case-insensitive by default)
[[tool.dbtective.manifest_tests]]
name = "no_select_star"
type = "has_forbidden_code"
forbidden_patterns = ["SELECT *"]
description = "Models should not use SELECT *."
# case_sensitive = false  # (optional, default)
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

# Case-sensitive match for exact patterns
[[tool.dbtective.manifest_tests]]
name = "no_hardcoded_schema"
type = "has_forbidden_code"
forbidden_patterns = ["raw_prod.", "analytics_prod."]
case_sensitive = true
severity = "warning"
description = "Use dbt selectors"

Relevant dbt code

-- models/clean_model.sql (PASS - no forbidden patterns)
SELECT
    id,
    name
FROM users
WHERE active = true

-- models/star_model.sql (FAIL - contains 'SELECT *')
SELECT * FROM users
-- Also matches: select * from users (case-insensitive by default)

Rule: `max_code_lines`

Manifest Rule

max_code_lines details

This rule enforces a maximum line count for dbt code objects, helping to maintain code readability and encourage modular design. Objects with empty code will also be flagged by this rule.

Configuration

type: Must be max_code_lines.
max_lines: (optional) The maximum number of lines allowed for the code. Defaults to 150.
applies_to: (optional) List of dbt object types to include.
- Default: ["models", "snapshots", "macros"]
- Options: models, snapshots, macros

Common Rule Config

name: Human-readable name of the rule.
severity: "error" (fail) or "warning" (warn only).
- (optional, defaults to "error" if not specified)
description: Human-readable explanation of the rule.
category: Override the default rule category. Included in structured output (JSON, CSV, NDJSON) but not in the CLI table. Each rule has a built-in default (e.g. documentation, naming, testing, governance, structure, performance).
- (optional, defaults to the rule type's built-in category)
includes: List of patterns to explicitly include for this rule. See Includes & Excludes for pattern syntax and examples.
excludes: List of patterns to explicitly exclude from this rule. See Includes & Excludes for pattern syntax and examples.
model_materializations: Filter models by materialization type. Only applies when applies_to includes models.
- (optional, if not specified all materializations are included)
- Built-in types: table, view, incremental, ephemeral, materialized_view. Custom materializations are also supported.
- Example: ["table", "incremental"]

Example Config

manifest_tests:
  - name: "models_max_100_lines"
    type: "max_code_lines"
    max_lines: 100
    description: "Models should not exceed 100 lines of code."
    # severity: "warning"  (optional)
    # applies_to: ['models', 'snapshots'] (optional)
    # includes: ["path/to/include/*"]
    # excludes: ["path/to/exclude/*"]

[[manifest_tests]]
name = "models_max_100_lines"
type = "max_code_lines"
max_lines = 100
description = "Models should not exceed 100 lines of code."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

[[tool.dbtective.manifest_tests]]
name = "models_max_100_lines"
type = "max_code_lines"
max_lines = 100
description = "Models should not exceed 100 lines of code."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

Relevant dbt code

-- models/short_model.sql (PASS)
SELECT
    id,
    name
FROM users

-- models/very_long_model.sql (FAIL - exceeds max_lines)
SELECT
    id,
    name,
    email,
    ...
    -- 101+ lines of SQL
    ...
FROM users

-- models/empty_model.sql (FAIL - empty code)
-- No content

Use cases

Enforce code modularity by limiting file size
Prevent overly complex transformations in single files
Encourage breaking down large models into smaller, reusable CTEs or models
Maintain consistent code readability across the project
Catch accidentally empty SQL files

Rule: `max_joins`

Manifest Rule

max_joins details

This rule enforces a maximum number of JOINs in raw SQL code, helping to reduce code complexity and encourage modular design. SQL comments (single-line `--` and multi-line `/* */`) are stripped before counting, so commented-out JOINs are not counted. Detection is case-insensitive.

Configuration

type: Must be max_joins.
max_joins: (optional) The maximum number of JOINs allowed. Defaults to 5.
applies_to: (optional) List of dbt object types to include.
- Default: ["models", "snapshots"]
- Options: models, snapshots, macros

Common Rule Config

name: Human-readable name of the rule.
severity: "error" (fail) or "warning" (warn only).
- (optional, defaults to "error" if not specified)
description: Human-readable explanation of the rule.
category: Override the default rule category. Included in structured output (JSON, CSV, NDJSON) but not in the CLI table. Each rule has a built-in default (e.g. documentation, naming, testing, governance, structure, performance).
- (optional, defaults to the rule type's built-in category)
includes: List of patterns to explicitly include for this rule. See Includes & Excludes for pattern syntax and examples.
excludes: List of patterns to explicitly exclude from this rule. See Includes & Excludes for pattern syntax and examples.
model_materializations: Filter models by materialization type. Only applies when applies_to includes models.
- (optional, if not specified all materializations are included)
- Built-in types: table, view, incremental, ephemeral, materialized_view. Custom materializations are also supported.
- Example: ["table", "incremental"]

Example Config

manifest_tests:
  - name: "limit_joins"
    type: "max_joins"
    max_joins: 3
    description: "Models should not exceed 3 JOINs."
    # severity: "warning"  (optional)
    # applies_to: ['models', 'snapshots'] (optional)
    # includes: ["path/to/include/*"]
    # excludes: ["path/to/exclude/*"]

[[manifest_tests]]
name = "limit_joins"
type = "max_joins"
max_joins = 3
description = "Models should not exceed 3 JOINs."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

[[tool.dbtective.manifest_tests]]
name = "limit_joins"
type = "max_joins"
max_joins = 3
description = "Models should not exceed 3 JOINs."
# severity = "warning"  # (optional)
# applies_to = ["models", "snapshots"]  # (optional)
# includes = ["path/to/include/*"]
# excludes = ["path/to/exclude/*"]

Relevant dbt code

-- models/simple_model.sql (PASS - 1 JOIN, within limit of 3)
SELECT
    a.id,
    b.name
FROM {{ ref('users') }} a
JOIN {{ ref('orders') }} b ON a.id = b.user_id

-- models/complex_model.sql (FAIL - 4 JOINs, exceeds limit of 3)
SELECT a.id
FROM {{ ref('users') }} a
JOIN {{ ref('orders') }} b ON a.id = b.user_id
JOIN {{ ref('products') }} c ON b.product_id = c.id
JOIN {{ ref('categories') }} d ON c.cat_id = d.id
JOIN {{ ref('suppliers') }} e ON d.sup_id = e.id

-- models/commented_model.sql (PASS - commented JOINs are not counted)
-- JOIN old_table ON ...
/* LEFT JOIN another_table ON ... */
SELECT a.id
FROM {{ ref('users') }} a
JOIN {{ ref('orders') }} b ON a.id = b.user_id

Rules columns (5)

code (4)

Rule: code_contains_refs

Rule: has_forbidden_code

Rule: max_code_lines

Rule: max_joins

Rule: `code_contains_refs`

Rule: `has_forbidden_code`

Rule: `max_code_lines`

Rule: `max_joins`