Column Dependencies¶

The synthetic data generator supports both intra-table (column-to-column) and inter-table dependencies, allowing you to create realistic data relationships.

Intra-table Dependencies (Column Dependencies)¶

Generate columns that depend on other columns within the same table using reference mappings.

Basic Example¶

columns:
  - name: "department"
    data_type: "string"
    nullable: false
    faker_function: "random_element"
    faker_options:
      elements: ["Engineering", "Sales", "Marketing"]

  - name: "job_level"
    data_type: "string"
    nullable: false
    depends_on: "department"
    reference_mapping:
      "Engineering": ["Senior Engineer", "Engineer", "Junior Engineer"]
      "Sales": ["Sales Manager", "Account Executive", "Sales Representative"]
      "Marketing": ["Marketing Director", "Marketing Manager", "Content Specialist"]

Inter-table Dependencies (Reference Tables)¶

Reference external tables for data values, enabling complex multi-table relationships.

Basic Reference Table¶

columns:
  - name: "country_code"
    data_type: "string"
    nullable: false
    reference_table:
      catalog: "reference_data"
      schema: "geographic"
      table: "countries"
      key_column: "code"

Dependent Reference Table¶

columns:
  - name: "country_code"
    data_type: "string"
    nullable: false
    reference_table:
      catalog: "reference_data"
      schema: "geographic"
      table: "countries"
      key_column: "code"

  - name: "city_name"
    data_type: "string"
    nullable: false
    depends_on: "country_code"
    reference_table:
      catalog: "reference_data"
      schema: "geographic"
      table: "cities"
      key_column: "name"  # Will filter cities where country_code matches parent value

Validation Rules¶

Dependent columns cannot be nullable
Parent columns must exist and be defined before dependent columns
Circular dependencies are detected and prevented (break after max depth of 10)
Either reference_mapping or reference_table is required for dependent columns
Cannot specify both reference_mapping and reference_table for the same column

Table Generation Order¶

When processing multiple configuration files, tables are automatically sorted by dependencies to ensure referenced tables exist before dependent tables are generated.