Athena-Express: Simplifying SQL queries on Amazon Athena

As published on the official AWS Partner Network Blog

Synopsis

Athena-Express can simplify executing SQL queries in Amazon Athena AND fetching cleaned-up JSON results in the same synchronous or asynchronous request - well suited for web applications.

Example:

Amazon Athena Background

Amazon Athena, launched at AWS re:Invent 2016, made it easier to analyze data in Amazon S3 using standard SQL. Under the covers, it uses Presto, which is an opensource SQL engine developed by Facebook in 2012 to query their 300 Petabyte data warehouse. It's incredibly powerful!

Amazon Athena combines the strength of Presto with serverless & self-managed capabilities of AWS. By simply pointing Athena to your data in Amazon S3, one could start querying using standard SQL. Most results are delivered within seconds and there’s no need for complex ETL jobs to prepare your data for analysis. This makes it easy for anyone with SQL skills to quickly analyze large-scale datasets.

How athena-express simplifies using Amazon Athena

athena-express simplifies integrating Amazon Athena with any Node.JS application - running as a standalone application or as a Lambda function. As a wrapper on AWS SDK, Athena-Express bundles the following steps listed on the official AWS Documentation:

1. Initiates a query execution
2. Keeps checking until the query has finished executing
3. Fetches the results of the query execution from Amazon S3

And as added features

4. Formats the results into a clean, user-friendly JSON array
5. Handles specific Athena errors by recursively retrying for ThrottlingException, NetworkingError, and TooManyRequestsException
6. Provides optional helpful stats including cost per query in USD
7. Fetching results (rows) via Pagination OR as a continuous stream
8. Synchrnous and Asynchornous fetching of results (rows)

Integrating with Amazon Athena without athena-express would require you to identify the appropriate API methods in the AWS SDK, stich them together sequentially, and then build out an error handling & retry mechanism for each of those methods.

athena-express can help you save time & effort in setting up this integration so that you can focus on core application development.

How is athena-express being used?

The most common use case is integrating a web front-end with Amazon Athena using athena-express as a backend. This backend could be any Node.JS application that could be hosted locally, or on an EC2 instance, or AWS Lambda.

Here is an example using AWS Lambda:

This architecture has a web front-end that invokes an API endpoint hosted on Amazon API Gateway by passing a query request. The query request can be as simple as SELECT * FROM movies LIMIT 3

This API Gateway then triggers a Lambda function that has the athena-express library imported.

Setup

Prerequisites

• You will need either an IAM Role (if you're running athena-express on AWS Lambda or AWS EC2) OR an IAM User with accessKeyId and secretAccessKey (if you're running athena-express on a standalone NodeJS application)
• This IAM role/user must have AmazonAthenaFullAccess and AmazonS3FullAccess policies attached
  • Note: As an alternative to granting AmazonS3FullAccess you could granularize and limit write access to a specific bucket. Just specify this bucket name during athena-express initialization

Configuration

• athena-express needs an AWS SDK object created with relevant permissions as mentioned in the prerequisites above.
• This AWS object is passed within the constructor so that it can invoke Amazon Athena SDK. It's up to you how you create this aws object. Here are 4 options:
  1. Create an aws object by explicitly passing in the accessKeyId and secretAccessKey generated in prerequisites

   const AWS = require("aws-sdk");
   const awsCredentials = {
       region: "YOUR_AWS_REGION",
   	accessKeyId: "YOUR_AWS_ACCESS_KEY_ID",
   	secretAccessKey: "YOUR_AWS_SECRET_ACCESS_KEY"
   };
   AWS.config.update(awsCredentials);
  
   const athenaExpressConfig = { aws: AWS }; //configuring athena-express with aws sdk object
   const athenaExpress = new AthenaExpress(athenaExpressConfig);

  2. OR if using Lambda, provide an IAM execution role with AmazonAthenaFullAccess and AmazonS3FullAccess policies attached

   const AWS = require("aws-sdk");
   const athenaExpressConfig = { aws: AWS }; //configuring athena-express with aws sdk object
   const athenaExpress = new AthenaExpress(athenaExpressConfig);

  3. OR Use instance profiles when using EC2s
  4. OR Use environment variables

Simple configuration

• Simple configuration requires only the AWS SDK object to be passed as a parameter to initialize athena-express
• Default values are assumed for all parameter options and athena-express creates a new S3 bucket in your AWS account for Amazon Athena to store the query results in.

const AWS = require("aws-sdk");

const athenaExpressConfig = { aws: AWS }; //simple configuration with just an aws sdk object

//Initializing athena-express
const athenaExpress = new AthenaExpress(athenaExpressConfig);

Advance configuration

• Besides the aws sdk paramater that is required, you can add any of the following optional parameters below

const AWS = require("aws-sdk");

//Example showing all Config parameters.
const athenaExpressConfig = {
	aws: AWS, // required 
	s3: "s3://mybucketname", // optional 
	db: "myDbName", // optional
	workgroup: "myWorkGroupName", // optional
	formatJson: true, // optional
	retry: 200, // optional
	getStats: true, // optional
	ignoreEmpty: true, // optional
	encryption: { EncryptionOption: "SSE_KMS", KmsKey: process.env.kmskey}, // optional
	skipResults: false, // optional
	waitForResults: false, // optional
	catalog: "hive" //optional
};

//Initializing AthenaExpress
const athenaExpress = new AthenaExpress(athenaExpressConfig);

Advance config Parameters:

Advance Query Parameters:

//Example showing all Query parameters.
let myQuery = {
    sql: "SELECT * FROM elb_logs LIMIT 3" // required,
    db: "sampledb", // optional. 
    pagination: 5, //optional
    NextToken: "ARfCDXRjMk...", //optional
    QueryExecutionId: "c274843b-4c5c-4ccf-ac8b-e33d595b927d", //optional
    catalog: "hive" //optional
};

Usage: Invoking athena-express

Using Promises to query Athena:

/*Option 1: object notation*/ 
let myQuery = {
	sql: "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3" /* required */,
	db: "sampledb" /* optional. You could specify a database here or in the advance configuration option mentioned above*/
};

/*OR Option 2: string notation*/ 
let myQuery = "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3"

athenaExpress
	.query(myQuery)
	.then(results => {
		console.log(results);
	})
	.catch(error => {
		console.log(error);
	});

Using Async/Await to query Athena:

(async () => {
/*Option 1: object notation*/ 
	let myQuery = {
		sql: "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3" /* required */,
		db: "sampledb" /* optional. You could specify a database here or in the configuration constructor*/
	};
    
/*OR Option 2: string notation*/ 
let myQuery = "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3"

	try {
		let results = await athenaExpress.query(myQuery);
		console.log(results);
	} catch (error) {
		console.log(error);
	}
})();

Using QueryExecutionID:

Applicable only if you already have the QueryExecutionID from an earlier execution. See skipResults or waitForResults in the advance config params above to learn more.

const myQueryExecutionId = "bf6ffb5f-6c36-4a66-8735-3be6275960ae";
let results = await athenaExpress.query(myQueryExecutionId);
console.log(results);

Full Examples

Using a standalone NodeJS application

"use strict";

const AthenaExpress = require("athena-express"),
	AWS = require("aws-sdk"),
	awsCredentials = {
		region: "YOUR_AWS_REGION",
		accessKeyId: "YOUR_AWS_ACCESS_KEY_ID",
		secretAccessKey: "YOUR_AWS_SECRET_ACCESS_KEY"
	};

AWS.config.update(awsCredentials);

const athenaExpressConfig = {
	aws: AWS,
	s3: "s3://my-bucket-for-storing-athena-results-us-east-1",
	getStats: true
};

const athenaExpress = new AthenaExpress(athenaExpressConfig);

//Invoking a query on Amazon Athena
(async () => {
	let myQuery = {
		sql: "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3",
		db: "sampledb"
	};

	try {
		let results = await athenaExpress.query(myQuery);
		console.log(results);
	} catch (error) {
		console.log(error);
	}
})();

Using AWS Lambda

"use strict";

const AthenaExpress = require("athena-express"),
	AWS = require("aws-sdk");

	/* AWS Credentials are not required here 
    /* Make sure the IAM Execution Role used by this Lambda 
    /* has the necessary permission to execute Athena queries 
    /* and store the result in Amazon S3 bucket
    /* See configuration section above under Setup for more info */

const athenaExpressConfig = {
	aws: AWS,
	db: "sampledb",
	getStats: true
};
const athenaExpress = new AthenaExpress(athenaExpressConfig);

exports.handler = async event => {
	const sqlQuery = "SELECT elb_name, request_port, request_ip FROM elb_logs LIMIT 3";

	try {
		let results = await athenaExpress.query(sqlQuery);
		return results;
	} catch (error) {
		return error;
	}
};

Results:

More Examples

Pagination

Query to fetch results (rows) for page 1

async function main() {
     const myQuery = {
        sql: "SELECT * from students LIMIT 100",
        pagination: 10
    };
    let results = await athenaExpress.query(myQuery);
    console.log(results);
}
main();

This will fetch the first 10 results (rows) off the 100 that exits in Athena. To query the next 10 rows, pass the values for NextToken and QueryExecutionId that were returned in the first query.

Query to fetch results (rows) for page 2 and beyond

async function main() {
     const myQuery = {
        sql: "SELECT * from students LIMIT 100",
        pagination: 10,
        NextToken: "ARfCDXRjMkQsR1NWziK1ARgiip3umf3q0/bZmNZWeQxUDc7iSToT7uJHy2yo8nL5FyxQoIIkuPh/zDD51xld7SoALA+zhMhpZg==",
        QueryExecutionId: "c274843b-4c5c-4ccf-ac8b-e33d595b927d",
    };
    let results = await athenaExpress.query(myQuery);
    console.log(results);
}
main();

UTILITY queries

Show Tables (single column result)

const results = await athenaExpress.query("SHOW TABLES");
console.log(results);

//Output:
{ Items:
  [ { row: 'default' },
    { row: 'sampledb' } ] }

Describe Table (dual column result)

const results = await athenaExpress.query("DESCRIBE elb_logs");
console.log(results);

//Output:
{ Items:
  [ { request_timestamp: 'string' },
    { elb_name: 'string' },
    { request_ip: 'string' },
    { request_port: 'int' },
    { backend_ip: 'string' },
    { backend_port: 'int' },
    { request_processing_time: 'double' },
    { backend_processing_time: 'double' },
    { client_response_time: 'double' },
    { elb_response_code: 'string' },
    { backend_response_code: 'string' },
    { received_bytes: 'bigint' },
    { sent_bytes: 'bigint' },
    { request_verb: 'string' },
    { url: 'string' },
    { protocol: 'string' },
    { user_agent: 'string' },
    { ssl_cipher: 'string' },
    { ssl_protocol: 'string' } ] }

Contributors

Gary Arora

License

MIT