Sometimes we need to test AWS Athena queries, but doing so directly can lead to unexpected costs if something goes wrong. To avoid this, you can test Athena-like queries locally using Trino. In this guide, we’ll install Trino on your laptop and run a simple query against a Parquet file.

Installing Trino:
I’m using Ubuntu 24.04 LTS. To get started, we’ll need to install Python and Java as prerequisites.

apt install python -y
sudo apt install openjdk-21-jdk -y

For confirmation:

python --version
java --version

Download Trino (Earlier known as Presto)

wget https://repo1.maven.org/maven2/io/trino/trino-server/438/trino-server-438.tar.gz

Extract the downloaded tar archive

tar -xvzf trino-server-438.tar.gz

Move it to the /opt directory (or any other location you prefer).

sudo mv trino-server-438 /opt/trino

Create configuration directory for Trino:

sudo mkdir -p /opt/trino/etc/catalog

Then create a config file:

sudo vim /opt/trino/etc/config.properties

Use the following content (feel free to modify the options, such as changing the port number, to suit your setup).

coordinator=true
node-scheduler.include-coordinator=true
http-server.http.port=8090
query.max-memory=5GB
query.max-memory-per-node=1GB
discovery-server.enabled=true
discovery.uri=http://localhost:8090

Create a jvm.config

sudo vim /opt/trino/etc/jvm.config

With the following content:

-server
-Xmx16G
-XX:+UseG1GC
-XX:+UseGCOverheadLimit
-XX:+ExplicitGCInvokesConcurrent
-XX:+HeapDumpOnOutOfMemoryError

Create node.properties

sudo vim /opt/trino/etc/node.properties

with the following content:

node.environment=production
node.id=trino-coordinator
node.data-dir=/opt/trino/data

For our demo, we will use Parquet files locally on same system where we are running Trino.
So we will create a hive catalogue like this:

sudo vim /opt/trino/etc/catalog/hive.properties

With the following content:

connector.name=hive
hive.metastore=file
hive.metastore.catalog.dir=file:///opt/trino/data/hive/warehouse
hive.metastore.user=trino

This configuration tells Trino to use /opt/trino/data/hive/warehouse as the data warehouse directory with a file-based metastore. We’ll place our Parquet file in this directory.

Now, we need to start the Trino server and connect to it. First, navigate to /opt/trino, then run the following command:

bin/launcher start

Now we need to install trino-cli ( Trino CLI provides a terminal-based, interactive shell for running queries)

wget https://repo1.maven.org/maven2/io/trino/trino-cli/438/trino-cli-438-executable.jar -O trino
chmod +x trino

To connect to Trino running on localhost, use the following command (we’re using port 8090, as specified earlier in the configuration file):

./trino --server localhost:8090 --catalog hive

I created a folder at:
/opt/trino/data/hive/warehouse/

sudo mkdir /opt/trino/data/hive/warehouse/my_table/

Then, I copied my Parquet file into that folder:

Next, we need to create a Hive table pointing to the folder containing our Parquet file. Be sure to define the column names and data types exactly as they appear in the Parquet file.

CREATE TABLE hive.default.mydata (
    fname VARCHAR,
    lastname VARCHAR
)
WITH (
    format = 'PARQUET',
    external_location = 'file:///opt/trino/data/hive/warehouse/my_table/'
);

Finally, we can run queries against our Parquet file.

SELECT * FROM hive.default.mydata WHERE fname = 'Karim';

I hope you find this helpful and that it allows you to test Athena queries freely without worry.