Skip to content

Latest commit

 

History

History
87 lines (56 loc) · 5.03 KB

README.md

File metadata and controls

87 lines (56 loc) · 5.03 KB
Raw

SimStreamer

NOTICE

  • This application is not part of the Snowflake Service and is governed by the terms in LICENSE file, unless expressly agreed to in writing. You use this application at your own risk, and Snowflake has no obligation to support your use of this application.

  • This utility is not production grade code rather than a quickly written sample to demonstrate a Snowflake functionality. It is not optimized for performance.

Problem Statement

Snowflake introduced Snowpipe Streaming API which allows to load streaming data rows in Snowflake at low-latency using the Snowflake Ingest SDK and your own managed application code. The streaming ingest API writes rows of data to Snowflake tables, unlike bulk data loads or Snowpipe, which write data from staged files. This architecture results in lower load latencies, with corresponding lower costs for loading similar volumes of data, which makes it a powerful tool for handling real-time data streams.

The streaming API can be leveraged either using the Snowflake Kafka Connector or a Java Snowflake Ingest SDK. This utility illustrates how to use the Java SDK. By using the utility, users are able to leverage this feature without having to setup a Kafka connector and connect to a streaming data source. All streaming data is is generated by the utility. The utility is multi-threaded allowing users to experience streaming multiple rowsets in parallel through multiple channels within a single client connection and observe the behavior.

Introduction

This utility leverages the Java Snowflake Ingest SDK to stream fake but realistic data into Snowflake. It's leveraging the Java Faker project for this purpose. Currently, the only use case supported is tpch data. It's a multi-threaded program which allows to configure the number of sessions to open to be able to load the data in parallel. Each thread will create a separate channel and stream the data in parallel in Snowflake.

NOTE: THIS IS NOT PRODUCTION GRADE CODE.

There are 2 ways to use this:

  • If you want to extend the utility to add your own use cases, you can pull the repo, add your use case and compile from source using maven
  • You could also use the jar executable IceStream.jar as-is for the existing use case.

Prerequisite

  • Clone this repository.
  • This has been developed using OpenJDK version 19.0.2. So you will need this JDK level:
java -version
openjdk version "19.0.2" 2023-01-17
OpenJDK Runtime Environment (build 19.0.2+7-44)
OpenJDK 64-Bit Server VM (build 19.0.2+7-44, mixed mode, sharing)

Use Cases

tpch Use Case

Snowflake comes by default with the TPC-H and TPC-DS sample datasets available at different scale factors within different schemas. For this utility, the choice was taken to use an existing sample schema so users can get to experiment some streaming use cases with a readily available schema definition and dataset within any Snowflake account.

Currently, the only use case supported. This will generate fake data in LINEITEM fact table, ensuring that the foreign keys are in the range of all the dimension tables. The current range is configured for SF (Scale Factor) 1. But, you could customize the range in simulation.json file.

  • Create a target lineitem table in a database of your choice as follows:
create table lineitem like snowflake_sample_data.tpch_sf1.lineitem;

Use compiled jar

  • You can use the 'SimStreamer.jar' compiled jar from the target subdirectory.
  • You need to customize the connection.jar file with properties pertaining to your Snowflake target account.
  • You can leave the simulation.json as-is, and customize the number of rows you want to stream and the number of channels you want to open to load rowsets in parallel.
  • You can now execute the jar as follows:
java -jar SimStreamer.jar --help
Unrecognized option: --help
usage: Usage:
 -c,--connection_parms <arg>   Path to connection.json file
 -s,--simulation_parms <arg>   Path to simulation.json file
 -u,--use-case <arg>           Use Cases supported: tpch
  • Once you have filled the connection.json & customized the simulation.json, you can run it as following:
java -jar SimStreamer.jar -c connection.json -s simulation.json -u tpch

Compile from source

  • Download this repo
  • Run the following mvn command from the directory where pom.xml is located:
mvn package
  • This will generate a SimStreamer-1.0-SNAPSHOT-jar-with-dependencies.jar under target directory that you can rename as SimStreamer.jar.