This Trino connector stores virtual views, which allow improvising, extending, and re-platforming data sources while seamlessly preserving compatability with all applications. Virtual views are particularly useful for prototyping and updating existing applications to use Apache Iceberg. This is a lightweight connector and does not require configuring a metastore or object storage.
This connector was originally published for my presentation on virtual view hierarchies at Trino Summit 2024, and is being maintained as a small way to give back to the Trino community.
- Java 23
- Trino 470
viewzoo.storage_type - Where the views will be stored, Either "filesystem" (default) or "jdbc".
# If viewzoo.storage_type=filesystem
viewzoo.dir
# If viewzoo.storage_type=jdbc
viewzoo.jdbc_url
viewzoo.jdbc_user
viewzoo.jdbc_password
1. Install Trino
download and expand tarball to local directory
export TRINO_HOME=$HOME/...
2. Create $TRINO_HOME/etc/catalog/viewzoo.properties:
connector.name=viewzoo
viewzoo.dir=/tmp/viewzoo
3. Build the connector and redeploy
mvn clean package && rm -rf $TRINO_HOME/plugin/viewzoo /tmp/viewzoo && cp -r ./target/viewzoo-470 $TRINO_HOME/plugin/viewzoo && mkdir -p /tmp/viewzoo
4. Start Trino
cd $TRINO_HOME
bash bin/launcher run
Create a view with a static result set:
create or replace view viewzoo.example.hello as select * from (values (1, 'a'), (2, 'b'), (3, 'c')) as t (key, value)Select rows from the view:
select * from viewzoo.example.helloDelete the view:
drop view viewzoo.example.hello© 2024-2025 Rob Dickinson (robfromboulder)