Sample Python3 example to run SQL queries concurrently

parallel-query.py

import concurrent.futures
import os

import trino


def execute_query(cur, query):
    cur.execute(query)
    return cur.fetchall()


# Create connection, only one of these is needed
conn = trino.dbapi.connect(
    host=os.getenv("TRINO_HOST", default="localhost"),
    port=os.getenv("TRINO_PORT", default="8080"),
    user=os.getenv("TRINO_USER"),
    catalog=os.getenv("TRINO_CATALOG"),
    schema=os.getenv("TRINO_SCHEMA"),
)

queries = ["SELECT * FROM system.runtime.nodes"] * 10

with concurrent.futures.ThreadPoolExecutor() as executor:
    futures = []
    for query in queries:
        futures.append(executor.submit(execute_query, cur=conn.cursor(), query=query))
    for future in concurrent.futures.as_completed(futures):
        print(future.result())

Trino Query Examples.md

Trino Query Examples

Python

• Install trino client as pip install trino.
• Set the env vars TRINO_HOST, TRINO_PORT, TRINO_USER, TRINO_CATALOG and TRINO_SCHEMA.
• Run as python3 <file.py>.

write-to-csv.py

import os

import trino


def write_results_as_csv(cur, query, csv_file):
    cur.arraysize = 10000
    cur.execute(query)
    with open(csv_file, 'w', newline='') as csvfile:
        writer = csv.writer(csvfile)
        rows = cur.fetchmany()
        # get column names
        headers = [desc[0] for desc in cur.description]
        write.writerow(headers)
        
        # write rows
        while len(rows) > 0:
            writer.writerows(rows)
            rows = cur.fetchmany()


# Create connection, only one of these is needed
conn = trino.dbapi.connect(
    host=os.getenv("TRINO_HOST", default="localhost"),
    port=os.getenv("TRINO_PORT", default="8080"),
    user=os.getenv("TRINO_USER"),
    catalog=os.getenv("TRINO_CATALOG"),
    schema=os.getenv("TRINO_SCHEMA"),
)

query = "SELECT * FROM system.runtime.nodes"
write_results_as_csv(conn.cursor(), query, 'results.csv')