Apache Gravitino Trino connector supported Catalogs

The catalogs currently supported by the Apache Gravitino Trino connector are as follows:

• Hive
• Iceberg
• MySQL
• PostgreSQL

Create catalog

Users can create catalogs through the Gravitino Trino connector and then load them into Trino. The Gravitino Trino connector provides the following stored procedures to create, delete, and alter catalogs. User can also use the system table catalog to describe all the catalogs.

Create catalog:

create_catalog(CATALOG varchar, PROVIDER varchar, PROPERTIES MAP(VARCHAR, VARCHAR), IGNORE_EXIST boolean);

• CATALOG: The catalog name to be created.
• PROVIDER: The catalog provider, currently only supports hive, lakehouse-iceberg, jdbc-mysql, jdbc-postgresql.
• PROPERTIES: The properties of the catalog.
• IGNORE_EXIST: The flag to ignore the error if the catalog already exists. It's optional, the default value is false.

The type of catalog properties reference:

• Hive catalog
• Iceberg catalog
• MySQL catalog
• PostgreSQL catalog

Drop catalog:

drop_catalog(CATALOG varchar, IGNORE_NOT_EXIST boolean);

• CATALOG: The catalog name to be deleted.
• IGNORE_NOT_EXIST: The flag to ignore the error if the catalog does not exist. It's optional, the default value is false.

Alter catalog:

alter_catalog(CATALOG varchar, SET_PROPERTIES MAP(VARCHAR, VARCHAR), REMOVE_PROPERTIES ARRY[VARCHAR]);

• CATALOG: The catalog name to be altered.
• SET_PROPERTIES: The properties to be set.
• REMOVE_PROPERTIES: The properties to be removed.

These stored procedures are under the gravitino connector and the system schema. So you need to use the following SQL to call them in the trino-cli:

Describe catalogs:

The system table gravitino.system.catalog is used to describe all the catalogs.

select * from gravitino.system.catalog;

The result is like:

     name     | provider |                                                 properties
--------------+----------+-------------------------------------------------------------------------------------------------------------
 gt_hive      | hive     | {gravitino.bypass.hive.metastore.client.capability.check=false, metastore.uris=thrift://trino-ci-hive:9083}

Example: You can run the following SQL to create a catalog named mysql with jdbc-mysql provider.

-- Call stored procedures with position.
call gravitino.system.create_catalog(
    'mysql',
    'jdbc-mysql',
    Map(
        Array['jdbc-url', 'jdbc-user', 'jdbc-password', 'jdbc-driver'],
        Array['jdbc:mysql://192.168.164.4:3306?useSSL=false', 'trino', 'ds123', 'com.mysql.cj.jdbc.Driver']
    )
);
call gravitino.system.drop_datalog('mysql');

-- Call stored procedures with name.
call gravitino.system.create_catalog(
    catalog =>'mysql',
    provider => 'jdbc-mysql',
    properties => Map(
        Array['jdbc-url', 'jdbc-user', 'jdbc-password', 'jdbc-driver'],
        Array['jdbc:mysql://192.168.164.4:3306?useSSL=false', 'trino', 'ds123', 'com.mysql.cj.jdbc.Driver']
    ),
    ignore_exist => true
);

call gravitino.system.drop_datalog(
    catalog => 'mysql'
    ignore_not_exist => true
);

call gravitino.system.alter_catalog(
    catalog => 'mysql',
    set_properties=> Map(
        Array['jdbc-url'],
        Array['jdbc:mysql://127.0.0.1:3306?useSSL=false']
    ),
    remove_properties => Array['jdbc-driver']
);

if you need more information about catalog, please refer to: Create a Catalog.

Passing Trino connector configuration

A Gravitino catalog is implemented by the Trino connector, so you can pass the Trino connector configuration to the Gravitino catalog. For example, you want to set the hive.config.resources configuration for the Hive catalog, you can pass the configuration to the Gravitino catalog like this:

call gravitino.system.create_catalog(
    'gt_hive',
    'hive',
    map(
        array['metastore.uris', 'trino.bypass.hive.config.resources'],
        array['thrift://trino-ci-hive:9083', '/tmp/hive-site.xml,/tmp/core-site.xml']
    )
);

A prefix with trino.bypass. in the configuration key is used to indicate Gravitino Trino connector to pass the Trino connector configuration to the Gravitino catalog in the Trino runtime.

More Trino connector configurations can refer to:

• Hive catalog
• Iceberg catalog
• MySQL catalog
• PostgreSQL catalog

Data type mapping between Trino and Apache Gravitino

Gravitino Trino connector supports the following data type conversions between Trino and Gravitino currently. Depending on the detailed catalog, Gravitino may not support some data types conversion for this specific catalog, for example, Hive does not support TIME data type.

Gravitino Type	Trino Type
Boolean	BOOLEAN
Byte	TINYINT
Short	SMALLINT
Integer	INTEGER
Long	BIGINT
Float	REAL
Double	DOUBLE
Decimal	DECIMAL
String	VARCHAR
Varchar	VARCHAR
FixedChar	CHAR
Binary	VARBINARY
Date	DATE
Time	TIME
Timestamp	TIMESTAMP
TimestampWithTimezone	TIMESTAMP WITH TIME ZONE
List	ARRAY
Map	MAP
Struct	ROW

For more about Trino data types, please refer to Trino data types and Gravitino data types, please refer to Gravitino data types.