Serverless plugin to load data on DynamoDB tables.
- Serverless version > 1.0.0. It has been tested with v1.36.3.
- Native support of promises: ECMAScript6 - Node 4.9.1
You can use this plugin in two ways:
- Via CLI: executing
sls fixtures
from your CLI. - Via deployment hook: after deploying your service (
sls deploy
). A hook is linked in this point (after deploying).
See the configuration flag enable
, for each fixture block, for more information.
You should include the plugin as dev dependency, and in your serverless.yml file. Then, as a custom variable, you should include a fixtures
variable with the configuration of the plugin. The fixtures configuration should contain a rules
element. and optionally an endpoint
element if you want to use a local instance of DynamoDB (remember to set the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables if you use dynamodb local).
The accepted rules configuration is an array of fixtures, where each fixture has the following variables:
- table: (String) Name of the DynamoDB table where you want to load your fixtures.
- enable: (String or boolean) Optional. This flag allows us to enable/disable the load of these fixtures. Accepted values:
- 'cli': enable the load of these fixtures only if the plugin was launched via CLI (
sls fixtures
). This is the default value. - 'deploy': enable the load of these fixtures only if the plugin was launched via deployment hook (
sls deploy
). - true: enable the load of these fixtures independently of how the plugin was launched.
- false: disable the load of these fixtures independently of how the plugin was launched.
- 'cli': enable the load of these fixtures only if the plugin was launched via CLI (
- sources: (Array) List of relative file paths where your data is, in a valid format for the function batchWrite of DynamoDB.DocumentClient.
- rawsources: (Array) List of relative file paths where your data is, in a valid format for the function batchWriteItem of DynamoDB.
- stage: (String) Optional. The name of the stage where you want to load these fixtures. If this variable is not set, the fixtures will be loaded for every stage.
- concurrentWrites: (Positive integer) Optional. The maximum number of concurrent writes sent to DynamoDB per source. DynamoDB limits to 25 the number of sent items per request, so we have to internally chunck the sources, but at the same time we can send concurrently these chuncks. With this variable you can control this aspect. The default value is 5.
- autogeneratedId: (String) Optional. Name of a key attribute in the table that this plugin should autogenate for you. Any attribute with the same name in the provided data files will be overwritten.
Important: sources
or rawsources
must be defined (at least one of them).
plugins:
- serverless-dynamodb-fixtures
custom:
fixtures:
endpoint: http://localhost:8000
rules:
- table: TABLE1-${self:custom.stage}
[enable: true]
sources:
- ./file1-${self:custom.stage}.yml
- ./file2-${self:custom.stage}.json
rawsources:
- ./rawFormatFile1-${self:custom.stage}.yml
- table: TABLE2-${self:custom.stage}
[stage: test]
[concurrentWrites: 5]
sources:
- ./file3-${self:custom.stage}.yml
serverless.yml
plugins:
- serverless-dynamodb-fixtures
custom:
fixtures:
- table: TABLE1-${self:custom.stage}
autogeneratedId: id
sources:
- ./file.yml
file.yml
- name: Jack London
- name: John Doe
It will be loaded in DynamoDB as follow:
- id: 45745c60-7b1a-11e8-9c9c-2d42b21b1a3e
name: Jack London
- id: 35625c60-8b1f-22e9-1c3c-1ae1c16a2f4e
name: John Doe
- id: 1
name: Jack London
- id: 2
name: John Doe
[
{"id":1, "name":"Jack London"},
{"id":2, "name":"John Doe"}
]
- How to develop the seeder: https://github.com/99xt/serverless-dynamodb-local
- Original idea: https://github.com/marc-hughes/serverless-dynamodb-fixtures-plugin