Skip to content

Hybrid Table

You can provide configuration to Snowflake Object Lifecycle Engine (SOLE) for the following operations with a hybrid table:

  • Manage the lifecycle of new and existing hybrid tables
  • Manage the grants of a hybrid table
  • Manage the tags applied to hybrid tables

Usage

- hybrid_table:
    name: <hybrid table-name>
    <configuration-key>: <value>
    database: rel(database.<database-name>)
    schema: rel(schema.<schema-name>)
    grants:
      <privilege>:
        - rel(role.<role-name>)
        - rel(role.<role-name>)
    shares:
      <privilege>:
        - rel(share.<share-name>)
        - rel(share.<share-name>)
    with_tags:
      <tag-name>:
        value: <tag-value>
        schema: <source-schema>
databases:
  <database-name>:
    schemas:
      <schema-name>:
        hybrid_tables:
          <hybrid table-name>:
            <configuration-key>: <value>
            grants:
              <privilege>:
                - <role-name>
                - <role-name>
            shares:
              <privilege>:
                - <share-name>
                - <share-name>
            with_tags:
              <tag-name>:
                value: <tag-value>
                schema: <source-schema>

Supported parameters

SOLE supports the parameters listed below.

Configuration Key Required/Optional Data Types and Values Description
columns Required Object: See here for a definition of columns Defines a column to create in the hybrid table. At least one is required.
comment Optional String Specifies a comment for the hybrid table
change_tracking Optional Boolean: false (default) Specifies whether to enable change tracking on the hybrid table
cluster_by Optional List of String A list of one or more hybrid table columns or expressions you can use as clustering key(s) for the hybrid table
data_retention_days Integer Integer: For standard edition 0 or 1, and for enterprise edition 0 to 90 (permanent hybrid tables) and 0 or 1 (transient hybrid table) Specifies the retention period for the hybrid table so that Time Travel actions (SELECT, CLONE, UNDROP) can be performed on historical data in the hybrid table. The default value is inherited from the parent schema.
deleted Optional Boolean: true enables deletion prevention, false does nothing Specifies what objects can be deleted
depends_on Optional List of String: See Object Dependencies for a definition of depends_on List of SOLE-managed objects that this hybrid table depends on. Objects defined in the list will be added to the generated HCL dependencies list to ensure proper creation order.
grants Optional Map: See Supported hybrid table grants to roles List of Privileges and Roles to which privileges are granted on the current hybrid table
manage_mode Optional String: all (default), none, grants Configures what properties to manage for the hybrid table.
See Changing Manage Mode before changing the value.
primary_key Optional Block Map (max 1): See here for a definition of primary_key Defines the primary key constraint to create on the hybrid table
unique Optional List of Block Maps: See here for a definition of unique Defines the unique constraints to create on the hybrid table
foreign_keys Optional List of Block Maps: See here for a definition of foreign_keys Defines the foreign key constraints to create on the hybrid table
shares Optional Map: See Supported hybrid table grants to shares List of Privileges and Roles to which privileges are granted on the current hybrid table
with_tags Optional Object: See here for a definition of with_tags List of tag references to apply to the current hybrid table

columns parameter

You can specify the columns of the hybrid table to manage.

List multiple columns in an object format in the columns parameter, each with the following supported parameters:

Configuration Key Required/Optional Data Types and Values Description
type Required String Data type of the Column
comment Optional String Comment on the column
default Optional Block List (max 1): See here for a definition of default Defines the column default value.
Note: Because of the limitations of Snowflake's ALTER TABLE ADD/MODIFY COLUMN, updates to default will not be applied.
identity Optional Block List (max 1): See here for a definition of identity Defines the identity start/step values for a column.
Note: identity and default are mutually exclusive.
nullable Optional Boolean Defines whether this column can contain null values.
Note: Depending on your Snowflake version, the default value will not suffice if this column is used in a primary key constraint.
with_tags Optional Object: See here for a definition of with_tags List of tag references to be applied to the current column

Example

databases:
  <database-name>:
    schemas:
      <schema-name>:
        hybrid_tables:
          <table-name>:
            columns:
              NAME:
                type: VARCHAR(16777216)
              YEAR:
                type: VARCHAR(16777216)
                comment: "Test column YEAR"
                nullable: true

default parameter

You should provide only one of the three keys.

Configuration Key Required/Optional Data Types and Values Description
constant Optional String/Boolean/Null/Number (Float or Integer) The default constant value for the column
expression Optional String The default expression value for the column
sequence Optional String The default sequence to use for the column
databases:
  <database-name>:
    schemas:
      <schema-name>:
        hybrid_tables:
          <table-name>:
            columns:
              column_1:
                Type: text
              COLUMN_2:
                type: "TIMESTAMP_NTZ"
                comment: "Test column 2"
                nullable: true
                default:
                  constant: "10"

identity parameter

Configuration Key Required/Optional Data Types and Values Description
start_num Optional Integer The number to start incrementing at
step_num Optional Integer Step size to increment by

Example

databases:
  <database-name>:
    schemas:
      <schema-name>:
        hybrid_tables:
          <table-name>:
            columns:
              NAME:
                identity:
                  start_num: 1
                  step_num: 5

primary_key parameter

Configuration Key Required/Optional Data Types and Values Description
keys Required List of String Columns to use in primary key
name Optional String Name of constraint

Example

databases:
  <database-name>:
    schemas:
      <schema-name>:
        hybrid_tables:
          <table-name>:
            primary_key:
              keys:
                - NAME
                - YEAR
            name: "PK_NAME_YEAR"

unique parameter

Info

The unique parameter is only supported in SOLE for data products.

Warning

Unique constraints for hybrid tables can only be added at the time of table creation. Hybrid tables cannot be altered to add constraints after creation. See Snowflake documentation on hybrid table constraints for more details.

You can define multiple unique constraints as a list of objects with the following parameters:

Configuration Key Required/Optional Data Types and Values Description
name Optional String Name of the unique constraint
keys Required List of String Columns to use in unique constraint

Example

unique:
  - name: "UQ_CUSTOMER_EMAIL"
    keys:
      - CUSTOMER_EMAIL
      - PREFERRED_CATEGORY
  - name: "UQ_PRODUCT_CODE"
    keys:
      - PRODUCT_CODE

foreign_keys parameter

Info

The foreign_keys parameter is only supported in SOLE for data products.

Warning

Foreign keys for hybrid tables can only be added at the time of table creation. Hybrid tables cannot be altered to add constraints after creation. See Snowflake documentation on hybrid table constraints for more details.

You can define multiple foreign key constraints as a list of objects with the following parameters:

Configuration Key Required/Optional Data Types and Values Description
name Optional String Name of the foreign key constraint
keys Required List of String Columns in the current table for the foreign key
referenced_table Required String Table that contains the referenced columns
referenced_keys Required List of String Columns in the referenced table

Example

foreign_keys:
  - name: "FK_CUSTOMER_PRODUCT"
    keys:
      - PREFERRED_PRODUCT_ID
    referenced_table: rel(hybrid_table.PRODUCT_CATALOG)
    referenced_keys:
      - PRODUCT_ID

with_tags parameter

You can specify a list of tags to apply to the hybrid table/column.

List multiple tags in the with_tags parameter in an object format, with the following parameters:

Configuration Key Required/Optional Data Types and Values Description
value Required String Value to assign to the tag
database Optional String Database of the Tag identifier. If omitted, it defaults to the name of the database the hybrid table/column is in.
schema Optional String Schema of the Tag identifier. If omitted, it defaults to the name of the schema the hybrid table/column is in.

Example

- hybrid_table:
    name: TABLE1
    database: rel(database.DB1)
    schema: rel(schema.SCHEMA1)
    with_tags:
      DEPARTMENT:
        value: Sales
        schema: TAG_LIBRARY
      CONFIDENTIALITY:
        value: Sensitive
        schema: TAG_LIBRARY
databases:
  DB1:
    schemas:
      SCHEMA1:
        hybrid_tables:
          TABLE1:
            with_tags:
              DEPARTMENT:
                value: Sales
                schema: TAG_LIBRARY
              CONFIDENTIALITY:
                value: Sensitive
                schema: TAG_LIBRARY

In-Place changes

Using data type aliases leads to in-place changes when running a PLAN or APPLY for hybrid tables.

Aliases such as TEXT are converted to their actual value VARCHAR(16777216) when an object is created in Snowflake. Further, the actual value is saved in the local state as well. Thus, when subsequent pipelines run, the engine assumes that a change has to be made to convert VARCHAR(16777216) to specified data-type alias e.g. TEXT.

To reduce the number of in-place changes, use the actual data type, such as VARCHAR(16777216) or NUMBER(38,0), instead of data-type aliases such as TEXT or INTEGER.

For a complete list of data types and their aliases, refer to the Snowflake data type summary.

Supported hybrid table grants to roles

Following are the privileges you can grant to roles in the hybrid table definition:

  • ALL PRIVILEGES
  • SELECT
  • INSERT
  • UPDATE
  • DELETE
  • TRUNCATE
  • REFERENCES
  • OWNERSHIP

ALL PRIVILEGES handling

When you define ALL PRIVILEGES in the SOLE configuration file, you grant all the privileges listed above to roles on this object except OWNERSHIP. However, the management of ALL PRIVILEGES in SOLE differs from its handling in Snowflake. See Handling ALL PRIVILEGES in SOLE for more information.

Supported hybrid table grants to shares

Following is the privilege you can grant to shares in the hybrid table definition:

  • SELECT

Inherited grants

You can define grants for a hybrid table in one of three sections:

  • Hybrid Table definition with parameter grants
  • Schema definition with parameter hybrid table_grants
  • Database definition with parameter hybrid table_grants

Local grants override grants defined in the parent's section. This means that hybrid table_grants defined in the database would be overridden by hybrid table_grants defined in the schema, and hybrid table_grants defined in the schema would be overridden by grants defined in the hybrid table.

If no overriding grants are defined, grants from parents are inherited for all hybrid tables. If hybrid table_grants is defined in the schema, all hybrid tables in the schema would inherit the grants defined. Similarly, if hybrid table_grants is defined in the database, all hybrid tables in all schemas would inherit the grants defined.

Example

Inheriting grants defined in a database

- database:
    name: SALES_RECORD
    hybrid_table_grants:
      UPDATE:
        - rel(role.HR_ROLE)
        - ACCOUNTADMIN
- schema:
    name: SALES
    database: rel(database.SALES_RECORD)
- hybrid_table:
    name: PRODUCT
    database: rel(database.SALES_RECORD)
    schema: rel(schema.SALES)
    columns:
      PRODUCT_ID:
        type: NUMBER(38,0)
      PRODUCT_NAME:
        type: VARCHAR(16777216)
    comment: "Product Info"
databases:
  SALES_RECORD:
    schemas:
      SALES:
        hybrid_tables:
          PRODUCT:
            columns:
              PRODUCT_ID:
                type: NUMBER(38,0)
              PRODUCT_NAME:
                type: VARCHAR(16777216)
            comment: "Product Info"
    hybrid_table_grants:
      UPDATE:
        - HR_ROLE
        - ACCOUNTADMIN

With this configuration, the hybrid table PRODUCT will have the following grants:

  • UPDATE to:
  • HR_ROLE
  • ACCOUNTADMIN

Inheriting grants defined in a schema

- database:
    name: SALES_RECORD
    hybrid_table_grants:
      UPDATE:
        - rel(role.HR_ROLE)
        - ACCOUNTADMIN
- schema:
    name: SALES
    database: rel(database.SALES_RECORD)
    hybrid_table_grants:
      INSERT:
        - DEV_ROLE
        - SYSADMIN
      TRUNCATE:
        - ACCOUNTADMIN
- hybrid_table:
    name: PRODUCT
    database: rel(database.SALES_RECORD)
    schema: rel(schema.SALES)
    columns:
      PRODUCT_ID:
        type: NUMBER(38,0)
      PRODUCT_NAME:
        type: VARCHAR(16777216)
    comment: "Product Info"
databases:
  SALES_RECORD:
    schemas:
      SALES:
        hybrid_table_grants:
          INSERT:
            - DEV_ROLE
            - SYSADMIN
          TRUNCATE:
            - ACCOUNTADMIN
        hybrid_tables:
          PRODUCT:
            columns:
              PRODUCT_ID:
                type: NUMBER(38,0)
              PRODUCT_NAME:
                type: VARCHAR(16777216)
            comment: "Product Info"
    hybrid_table_grants:
      UPDATE:
        - HR_ROLE
        - ACCOUNTADMIN

With this configuration, the hybrid table PRODUCT will have the following grants:

  • INSERT to:
  • DEV_ROLE
  • SYSADMIN
  • TRUNCATE to:
  • ACCOUNTADMIN

Overriding grants in a hybrid table

- database:
    name: SALES_RECORD
    hybrid_table_grants:
      UPDATE:
        - rel(role.HR_ROLE)
        - ACCOUNTADMIN
- schema:
    name: SALES
    database: rel(database.SALES_RECORD)
    hybrid_table_grants:
      INSERT:
        - DEV_ROLE
        - SYSADMIN
      TRUNCATE:
        - ACCOUNTADMIN
- hybrid_table:
    name: PRODUCT
    database: rel(database.SALES_RECORD)
    schema: rel(schema.SALES)
    grants:
      SELECT:
        - rel(role.DEV_ROLE)
      UPDATE:
        - rel(role.DEV_ROLE)
    columns:
      PRODUCT_ID:
        type: NUMBER(38,0)
      PRODUCT_NAME:
        type: VARCHAR(16777216)
    comment: "Product Info"
databases:
  SALES_RECORD:
    schemas:
      SALES:
        hybrid_table_grants:
          INSERT:
            - DEV_ROLE
            - SYSADMIN
          TRUNCATE:
            - ACCOUNTADMIN
        hybrid_tables:
          PRODUCT:
            grants:
              SELECT:
                - DEV_ROLE
              UPDATE:
                - DEV_ROLE
            columns:
              PRODUCT_ID:
                type: NUMBER(38,0)
              PRODUCT_NAME:
                type: VARCHAR(16777216)
            comment: "Product Info"
    hybrid table_grants:
      UPDATE:
        - HR_ROLE
        - ACCOUNTADMIN

With this configuration, the hybrid table PRODUCT will have the following grants:

  • SELECT to:
  • DEV_ROLE
  • UPDATE to:
  • DEV_ROLE

Examples

- hybrid_table:
    name: PRODUCT
    database: rel(database.SALES_RECORD)
    schema: rel(schema.SALES)
    columns:
      PRODUCT_ID:
        type: NUMBER(38,0)
      PRODUCT_NAME:
        type: VARCHAR(16777216)
      PURCHASE_DATE:
        type: TIMESTAMP_NTZ(9)
        comment: "Product purchase date"
        nullable: true
        default:
          expression: "CURRENT_TIMESTAMP()"
      SERIAL:
        type: NUMBER(38,0)
        identity:
          start_num: 1
          step_num: 2
    comment: "Product Info"
    data_retention_days: 10
    change_tracking: false
    primary_key:
      keys:
        - PRODUCT_ID
      name: PK_PRODUCT_ID_SERIAL
    cluster_by:
      - "to_date(PURCHASE_DATE)"
      - "PRODUCT_NAME"
    grants:
      SELECT:
        - rel(role.DEV_ROLE)
      UPDATE:
        - rel(role.DEV_ROLE)
databases:
  SALES_RECORD:
    schemas:
      SALES:
        hybrid_tables:
          PRODUCT:
            columns:
              PRODUCT_ID:
                type: NUMBER(38,0)
              PRODUCT_NAME:
                type: VARCHAR(16777216)
              PURCHASE_DATE:
                type: TIMESTAMP_NTZ(9)
                comment: "Product purchase date"
                nullable: true
                default:
                  expression: "CURRENT_TIMESTAMP()"
              SERIAL:
                type: NUMBER(38,0)
                identity:
                  start_num: 1
                  step_num: 2
            comment: "Product Info"
            data_retention_days: 10
            change_tracking: false
            primary_key:
              keys:
                - PRODUCT_ID
              name: PK_PRODUCT_ID_SERIAL
            cluster_by:
              - "to_date(PURCHASE_DATE)"
              - "PRODUCT_NAME"
            grants:
              SELECT:
                - DEV_ROLE
              UPDATE:
                - DEV_ROLE

Examples with Constraints

- hybrid_table:
    name: CUSTOMER_DATA
    database: rel(database.SALES_RECORD)
    schema: rel(schema.SALES)
    columns:
      CUSTOMER_ID:
        type: NUMBER(38,0)
        nullable: false
      CUSTOMER_NAME:
        type: VARCHAR(100)
        nullable: false
      CUSTOMER_EMAIL:
        type: VARCHAR(255)
        nullable: false
      PREFERRED_PRODUCT_ID:
        type: NUMBER(38,0)
        nullable: true
    comment: "Customer information"
    primary_key:
      keys:
        - CUSTOMER_ID
      name: PK_CUSTOMER
    unique:
      name: UQ_CUSTOMER_EMAIL
      keys:
        - CUSTOMER_EMAIL
    foreign_keys:
      - name: FK_CUSTOMER_PRODUCT
        keys:
          - PREFERRED_PRODUCT_ID
        referenced_table: rel(hybrid_table.PRODUCT_CATALOG)
        referenced_keys:
          - PRODUCT_ID
databases:
  SALES_RECORD:
    schemas:
      SALES:
        hybrid_tables:
          CUSTOMER_DATA:
            columns:
              CUSTOMER_ID:
                type: NUMBER(38,0)
                nullable: false
              CUSTOMER_NAME:
                type: VARCHAR(100)
                nullable: false
              CUSTOMER_EMAIL:
                type: VARCHAR(255)
                nullable: false
              PREFERRED_PRODUCT_ID:
                type: NUMBER(38,0)
                nullable: true
            comment: "Customer information"
            primary_key:
              keys:
                - CUSTOMER_ID
              name: PK_CUSTOMER
            unique:
              name: UQ_CUSTOMER_EMAIL
              keys:
                - CUSTOMER_EMAIL
            foreign_keys:
              - name: FK_CUSTOMER_PRODUCT
                keys:
                  - PREFERRED_PRODUCT_ID
                referenced_table: PRODUCT_CATALOG
                referenced_keys:
                  - PRODUCT_ID