Object Storage Service:Use Trino on an EMR cluster to query data stored in OSS-HDFS

Prerequisites

Before you begin, make sure that you have:

• An EMR cluster of EMR V3.42.0 or later, or EMR V5.8.0 or later, with the Trino service selected. For more information, see Create a cluster.

• A bucket with OSS-HDFS enabled and the required access permissions granted. For more information, see Enable OSS-HDFS and grant access permissions.

Step 1: Create an EMR cluster

1. Log on to the EMR console. In the left-side navigation pane, click EMR on ECS.

2. Create an EMR cluster with the following settings:

   Parameter	Required value
   Product Version	EMR-3.46.2 or later, or EMR-5.12.2 or later
   Root Storage Directory of Cluster	A bucket with OSS-HDFS enabled
   Other parameters	Default values

   For more information, see Create a cluster.

Step 2: Connect to the Trino server

1. Get the Trino server address and port:

   1. On the EMR on ECS page, click the name of your cluster.

   2. Go to Services > Trino > Configure to find the server address and port.

2. Run the following command to connect to the Trino server:

   Parameter	Description
   <Trino_server_address>	The IP address or hostname of the Trino server. Located on the Configure tab.
   <Trino_server_port>	The port number of the Trino server. Located on the Configure tab.
   <catalog_name>	The catalog to connect to, such as hive.

      trino --server <Trino_server_address>:<Trino_server_port> --catalog <catalog_name>

   Note: The --catalog flag requires a value. Specify the catalog that maps to your data source.

Step 3: Query data stored in OSS-HDFS

After you connect to the Trino server, run the following SQL statements to create a schema, load data, and run a query.

1. Create a schema that points to an OSS-HDFS location.

      create schema testDB with (location='oss://<Bucket>.<Endpoint>/<schema_dir>');

   Replace the placeholders with your values:

   Placeholder	Description	Example
   <Bucket>	The name of the OSS bucket with OSS-HDFS enabled	my-data-bucket
   <Endpoint>	The OSS-HDFS endpoint for your region	cn-hangzhou.oss-dls.aliyuncs.com
   <schema_dir>	The directory path for the schema	trino/testDB

   The full URI follows this format:

      oss://<Bucket>.<Endpoint>/<schema_dir>

2. Switch to the new schema.

      use testDB;

3. Create a table.

      create table tbl (key int, val int);

4. Insert data into the table.

      insert into tbl values (1,666);

5. Query data in the table.

      select * from tbl;

   Expected output:

       key | val
      -----+-----
         1 | 666
      (1 row)

Result

The query returns the row you inserted. The data is stored in the OSS-HDFS location specified in the schema.

References

• Create a cluster

• Enable OSS-HDFS and grant access permissions