Skip to content

Latest commit

 

History

History
97 lines (59 loc) · 2.07 KB

README.md

File metadata and controls

97 lines (59 loc) · 2.07 KB
Raw

lambda-dynamodb-bigquery

AWS Lambda function for insert DynamoDB records to Google BigQuery

How to build

This package depends on gcloud and it uses native code. So you have to build on Amazon Linux to execute AWS Lambda.

  • Launch new Amazon Linux instance

  • Install required packages to the instance

sudo yum update
sudo yum install git-core
sudo yum groupinstall "Development Tools"
  • Install nvm
curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.31.0/install.sh | bash
  • Install Node.js
nvm install v0.10.44
  • Clone repository
git clone https://github.com/runtakun/lambda-dynamodb-bigquery.git
  • Change directory and install npm packages
cd lambda-dynamodb-bigquery
npm install

  • Add configuration file and account key file

  • Install grunt and build package

npm install -g grunt-cli
grunt lambda_package
  • Create AWS Lambda function and upload package

Create Lambda function in AWS console and speficy Node.js 0.10 in Runtime section.

Account key

Create service account at developer console (below link) and download key file.

https://console.cloud.google.com/permissions/serviceaccounts

Configuration

You should specify project name and dataset id by JSON file named gcpconfig.json.

Example:

{"project": "lambda-bigquery-sample", "dataset": "sample"}

Table name

By default, BigQuery table will be named with the same as DynamoDB's one. But you can specify specific name by configuration.

Example:

{"project": "lambda-bigquery-sample", "dataset": "sample", "table": "Sample"}

Table partitioning

You can use table partitioning by configuration field tablePartitionPeriod and specifing daily or monthly. You should create normal table before data is inserted. For example, if you creates table named Sample on BigQuery, BigQuery creates Sample20160401 table.

Example:

{"project": "lambda-bigquery-sample", "dataset": "sample", "tablePartitionPeriod": "monthly"}

THANKS