Syntax

Reference documentation for Veza Query Language.

This document provides information about the VQL (Veza Query Language) query specification, which offers a SQL-like interface to the Assessment Query API. See the Quick Start Guide for detailed usage and examples.

• Concepts

• VQL Expression Syntax

• NodeSpec Syntax

• Usage Examples

  • Where Clause Operators

    • Comparison Operators

  • Data Types and Restrictions

  • Entity Attributes

• Permissions

  • System Permissions

  • Effective Permissions

• Intermediate Nodes

• Modifiers and Filters

• Result Output

Concepts

Early Access: VQL is subject to modifications as we add and improve functionality. Future updates will aim to preserve compatibility with earlier versions.

A VQL query is composed of the following components:

• Source Nodes: Node types in VQL represent entity types within Veza's Identity Graph. Each node type can have many individual instances, returned as rows in the query output.

• Path requirements: Graph nodes can be interrelated, forming complex graph structures. Specifying related nodes filters results with a matching relationship.

• Filters and Modifiers: Filter expressions (WHERE clauses) to constrain results based on attributes or other criteria.

VQL Expression Syntax

A VQL query always includes a SHOW statement describing the source node type. The general syntax is:

SHOW [NodeSpec] -- Specifies the source node type
[ [NOT] RELATED TO [NodeSpec]] -- Filters results by related node type
[WITH | NOT WITH] PATH [NodeSpec] -- Filters on intermediate nodes
[WHERE (filter expressions)]
[RESULT INCLUDE [DESTINATION NODES | DESTINATION NODE COUNT | PATH SUMMARY]]
[HAVING [entity_result_count (condition) | percentage_of_total_count (condition)]]
[WITH QUERY OPTIONS (options)];

NodeSpec Syntax

A NodeSpec describes a source or destination entity type. It can include attribute selection (which columns to return) and filters using a WHERE clause. The full syntax is:

[NodeType] [{ attribute1, attribute2, ... }] [WHERE (<attribute_name> <operator> <value> [AND|OR] ...)]

Basic Components:

• NodeType: The type of node (e.g., AwsIamUser, OktaGroup)

• Attribute Selection: Optional curly braces { } containing a comma-separated list of attributes to include in the results

• WHERE: Optional clause that applies attribute-based filters to the nodes using the selected operators

Usage Examples

Basic node specification (returns all attributes):

SHOW AwsIamUser

Select specific attributes to display:

SHOW AwsIamUser { created_at, is_active, full_admin }

Apply filters without attribute selection:

SHOW AwsIamUser WHERE is_active = true

Combine attribute selection and filters:

SHOW AwsIamUser { created_at, email } WHERE is_active = true AND risk_score > 70

In this more complex example, we select specific attributes for both the source (OktaUser) and destination (S3Bucket) nodes while also applying filters to the results:

SHOW OktaUser { email, last_login_at, department }
WHERE last_login_at >= CURRENT_DATE - 30
RELATED TO S3Bucket
WHERE block_public_acls = false
WITH PATH OktaGroup
WITH QUERY OPTIONS ( over_provisioned_score > 85 )

Where Clause Operators

A range of operators can be used to filter results depending on node properties. Currently, VQL supports the comparison operators:

Comparison Operators

Operator	Supported Data Types	Example
<, >, <=, >=, =, !=	NUMERIC, TIMESTAMP, TIME FUNCTIONS	risk_score < 80 created_at >= '2023-10-05 14:30:00.123' created_at < CURRENT_DATE - 30
STARTS_WITH	STRING	name STARTS_WITH 'S'
ENDS_WITH	STRING	name ENDS_WITH 'E'
LIST_CONTAINS	STRING	permissions LIST_CONTAINS 'iam:PassRole'
LIST_ALL_ELEMENTS_IN	STRING	accounts_assumed_by LIST_ALL_ELEMENTS_IN ('accountid1', 'accountid2')
REGEX	STRING	name REGEX 'TEst.*'
LIST_ANY_ELEMENT_EQ	STRING	permissions LIST_ANY_ELEMENT_EQ 'iam:SetDefaultPolicyVersion'
LIST_ANY_ELEMENT_STARTS_WITH	STRING	cai_tags LIST_ANY_ELEMENT_STARTS_WITH 'P'
LIST_ANY_ELEMENT_CONTAINS	STRING	cai_tags LIST_ANY_ELEMENT_CONTAINS 'policy'
LIST_ANY_ELEMENT_ENDS_WITH	STRING	cai_tags LIST_ANY_ELEMENT_ENDS_WITH 'admin'
LIST_ANY_ELEMENT_REGEX	STRING	cai_tags LIST_ANY_ELEMENT_REGEX '::'
IS NULL	STRING	show Key WHERE last_rotated_at IS NULL
IS NOT NULL	STRING	show Key WHERE last_rotated_at IS NOT NULL

Data Types and Restrictions

• Data Types: VQL supports boolean, integer, string, and null data types.

• Case Sensitivity: VQL is typically case-sensitive for:

  • Node Types: Must be written exactly as defined (e.g., AwsIamUser, not awsiamuser).

  • Attribute Names: Must match the exact casing (e.g., is_active, not Is_Active).

Entity Attributes

Entity attributes in VQL are used to filter and select graph nodes. They consist solely of alphanumeric characters or underscores (e.g., last_login, email_address).

Permissions

In addition to attribute filters, VQL queries can use permission filters. Both system permissions and effective permissions are supported.

System Permissions

System permissions are raw, system-level permissions and vary depending on the specific integration and resource. The following query identifies users that specifically have the ability to create new S3 buckets:

SHOW AwsIamUser
RELATED TO S3Bucket
WITH SYSTEM PERMISSIONS = ANY ('s3:CreateBucket');

• ANY: Used to filter for any of the supplied permissions (logical OR).

• ALL: Ensures that all permissions passed should be present on the resource (logical AND).

Effective Permissions

Veza supports nine different effective permissions. These are abstracted permissions that express system permissions in common groups:

• METADATA_READ

• METADATA_WRITE

• METADATA_CREATE

• METADATA_DELETE

• DATA_READ

• DATA_WRITE

• DATA_CREATE

• DATA_DELETE

• NON_DATA

The following query shows all AWS IAM Roles that grant effective permissions to read or write S3 bucket metadata:

SHOW AwsIamRole
RELATED TO S3Bucket
WHERE is_active = false
WITH EFFECTIVE PERMISSIONS = ALL ('METADATA_READ', 'METADATA_WRITE');

Effective permissions filters require a destination type (specified by the RELATED TO clause).

Intermediate Nodes

Intermediate node options include or exclude results based on certain node types within the path, for analyzing complex relationships that involve hierarchies of groups, roles, or entities. These queries are often used to find users whose access is (or is not) granted by group membership or role assignment:

• Including Intermediate Nodes:

  SHOW OktaUser
  RELATED TO SnowflakeTable
  WITH PATH SnowflakeUser;

• Excluding Intermediate Nodes:

  SHOW OktaUser
  RELATED TO SnowflakeTable
  NOT WITH PATH OktaGroup;

Modifiers and Filters

Use the WHERE clause to apply filters. You can combine multiple conditions with AND or OR statements:

Example:

SHOW AwsIamUser
WHERE is_active = true AND risk_score > 70;

Result Output

By default, queries return a list of source entity types and attributes. Use the RESULT INCLUDE clause to specify how query results appear:

• DESTINATION NODES: Includes information about related destination nodes.

• DESTINATION NODE COUNT: Provides a count of related destination nodes.

• PATH SUMMARY: Summarizes the paths between source and destination nodes.

Example:

SHOW AwsIamUser
RELATED TO S3Bucket
RESULT INCLUDE DESTINATION NODES;