Moving data from DynamoDB to Cloudant or CouchDB

Introducing the dynamodbexport npm module

If you have data in an Amazon DynamoDB service and want to move it to IBM Cloudant or Apache CouchDB, how would you go about it? First of all, DynamoDB has a peculiar form of JSON. A single temperature measurement would be expressed like this:

"temperature": {
"N": "8391"
"time": {
"S": "2017-03-09T01:38:11+0000"
"id": {
"S": "1489023491"

Instead of the more straightforward JSON:

"temperature": 8391,
"time": "2017-03-09T01:38:11+0000",
"id": "1489023491"

Cloudant and CouchDB can store any JSON documents with nested objects of arbitrary complexity, whereas DynamoDB stores a series of key values at the top of the JSON tree.

Getting the data out

The AWS SDK provides a comprehensive toolkit for interacting with AWS services. You need to to “scan” a whole DynamoDB table, performing a chain of API calls to pull back records in batches until you’ve consumed them all. I’ve written a script to do this for you: dynamodbexport.

The aptly-named dynamodbexport npm module moves data from DynamoDB to IBM Cloudant or Apache CouchDB™.

First, install the tool:

$ npm install -g dynamodbexport

Define a couple of environment variables with your Amazon API credentials:


Then simply run dynamodbexport, supplying the name of the table to export and the AWS region it is hosted in:

$ dynamodbexport --table iot --region us-east-1
Export complete { iterations: 1, records: 3, time: 0.145 }

The tool makes as many API calls as it needs to extract the data, converting the JSON to a more compact form as it goes.

Importing into CouchDB/Cloudant

I already have a tool to import data into CouchDB: couchimport, which you can install in a similar way:

$ npm install -g couchimport

Set an environment variable with your target Cloudant/CouchDB service’s URL:

$ export COUCH_URL=""

Then run both the dynamodbexport and couchimport commands together, piping the output of the former into the latter:

$ dynamodbexport --table iot --region us-east-1 | couchimport --db iot --type jsonl

The --type jsonl parameter tells couchimport that it is to expect one JSON document per line, and --db iot defines the name of the target database. (Make sure your target database already exists, since couchimport does not create new databases.)

It’s that simple! You can now use Cloudant’s awesome MapReduce tools to aggregate the data or replicate it to other devices.

It works for local databases too

The dynamodbexport tool also works for local DynamoDB databases. Just leave out the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables, and it will assume a local database.

Moving data from local DynamoDB to local CouchDB is as simple as:

$ dynamodbexport --table iot | couchimport --db iot

You’ll find more details on command-line usage and programmatic access for dynamodbexport on npm. And please ♡ this article if you’d like to recommend it to other Medium readers.