Package Exports

dynameh

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (dynameh) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Dynameh

DynamoDB on Node more easier

Dynameh makes the official DynamoDB JavaScript API easier to use. It makes it easier to:

Installation

Dynameh is your typical NPM package.

npm install --save dynameh

Usage

import * as dynameh from "dynameh";
// or
const dynameh = require("dynameh");

See the documentation for details on each module and its methods.

A Simple Example

This example is written using async/await which is available in TypeScript and ES2017.

import * as aws from "aws-sdk";
import * as dynameh from "dynameh";

// Initialize the DynamoDB client.
const dynamodb = new aws.DynamoDB({
    apiVersion: "2012-08-10",
    region: "us-west-1"
});

// Set up the table schema.
const tableSchema = {
    tableName: "motorcycles",
    partitionKeyField: "id",
    partitionKeyType: "string"
};

async function updateMotorcycleHorsePower(motorcycleId, bhp) {
    // Fetch the item from the database.
    const getRequest = dynameh.requestBuilder.buildGetInput(tableSchema, motorcycleId);
    const getResult = await dynamodb.getItem(getRequest).promise();
    let motorcycle = dynameh.responseUnwrapper.unwrapGetOutput(getResult);
    
    if (!motorcycle) {
        // Item not found, create it.
        motorcycle = {
            id: motorcycleId
        };
    }
    
    // Update the horse power stat.
    motorcycle.bhp = bhp;
    
    // Put the updated object in the database.
    const putRequest = dynameh.requestBuilder.buildPutInput(tableSchema, motorcycle);
    await dynamodb.putItem(putRequest).promise();
}

updateMotorcycleHorsePower("sv-650", 73.4);

TableSchema

The key to easy building of requests in Dynameh is the TableSchema. This simple object defines all the extra information Dynameh needs to build requests.

For a table called MyTable with a partition key id that is a number...

{
  "tableName": "MyTable",
  "partitionKeyField": "id",
  "partitionKeyType": "number"
}

For a table called MyAdvancedTable with a partition key id that is a string, a sort key date that is a number, and a version field version...

{
  "tableName": "MyAdvancedTable",
  "partitionKeyField": "id",
  "partitionKeyType": "string",
  "sortKeyField": "date",
  "sortKeyType": "number",
  "versionKeyField": "version"
}

Optimistic Locking

Optimistic locking is a strategy for preventing changes from clobbering each other. For example two processes read from the database, make unrelated changes, and then both write to the database but the second write overwrites the first (clobbers).

Enable optimistic locking by setting the versionKeyField on your TableSchema. In the second TableSchema example that field is version. The versionKeyField will be automatically incremented on the server side during a put request. If the value for versionKeyField sent does not match the current value in the database then the contents have changed since the last get and the optimistic lock has failed. In that case you should get the latest version from the database and replay the update against that.

import * as aws from "aws-sdk";
import * as dynameh from "dynameh";

// Initialize the DynamoDB client.
const dynamodb = new aws.DynamoDB({
    apiVersion: "2012-08-10",
    region: "us-west-1"
});

// Set up the table schema.
const tableSchema = {
    tableName: "motorcycles",
    partitionKeyField: "id",
    partitionKeyType: "string",
    versionKeyField: "version"
};

async function updateMotorcycleHorsePower(motorcycleId, bhp) {
    // Fetch the item from the database.
    const getRequest = dynameh.requestBuilder.buildGetInput(tableSchema, motorcycleId);
    const getResult = await dynamodb.getItem(getRequest).promise();
    let motorcycle = dynameh.responseUnwrapper.unwrapGetOutput(getResult);
    
    if (!motorcycle) {
        // Item not found, create it.
        // Note that we don't need to set the version on create.
        motorcycle = {
            id: motorcycleId
        };
    }
    
    // Update the horse power stat.
    motorcycle.bhp = bhp;
    
    // Put the updated object in the database.
    const putRequest = dynameh.requestBuilder.buildPutInput(tableSchema, motorcycle);
    try {
        await dynamodb.putItem(putRequest).promise();
    } catch (err) {
        if (err.code === "ConditionalCheckFailedException") {
            // If this is the error code then the optimistic locking has failed
            // and we should redo the update operation (done here with recursion).
            updateMotorcycleHorsePower(motorcycleId, bhp);
        } else {
            throw err;
        }
    }
}

updateMotorcycleHorsePower("sv-650", 73.4);

Query and Scan

Both query and scan are paginated operations. They may return a limited set of results and you must make repeated calls to get all the results. queryHelper and scanHelper provide utilities that go through the pagination for you.

queryAll() and scanAll() will collect all results in a single array. For small result sets this is easiest to work with but for large result sets this may require a lot of memory to store all results at once.

queryByCallback() and scanByCallback() take a callback parameter that processes each page of results. Processing one page at a time reduces the amount of memory required.

This example function scans the table for items and deletes them all one at a time:

async function deleteAllItems(dynamodbClient, tableSchema) {
    const scanInput = buildScanInput(tableSchema);
    await scanByCallback(dynamodbClient, scanInput, async items => {
        for (const item of items) {
            const delInput = buildDeleteInput(tableSchema, item);
            await dynamodbClient.deleteItem(delInput).promise();
        }
        return true;    // return true to continue, false to stop pagination
    });
}

Conditions

Conditions can be added to a put or delete request to make the operation conditional.

One of the most useful conditions is that the item must not already exist (create but not update). This is done by asserting attribute_not_exists on the primary key. For example...

const tableSchema = {
    tableName: "Boats",
    partitionKeyField: "name",
    partitionKeyType: "string"
};

async function addNewBoat(boat) {
    const putRequest = dynameh.requestBuilder.buildPutInput(tableSchema, boat);
    dynameh.requestBuilder.addCondition(tableSchema, putRequest, {attribute: "name", operator: "attribute_not_exists"});
    
    try {
        await dynamodb.putItem(putRequest).promise();
    } catch (err) {
        if (err.code === "ConditionalCheckFailedException") {
            throw new Error("This boat already exists!");
        } else {
            throw err;
        }
    }
}

addNewBoat({
    name: "Boaty McBoatface",
    type: "submarine",
    autonomous: true
});

The following conditions are available...

condition	# of value parameters	description
=	1	the attribute's value equals the supplied value
<>	1	the attribute's value does not equal the supplied value
<	1	the attribute's value is less than the supplied value
<=	1	the attribute's value is less than or equal to the supplied value
>	1	the attribute's value is greater than the supplied value
>=	1	the attribute's value is greater than or equal to the supplied value
BETWEEN	2	the attribute's value is between the supplied values
IN	at least 1	the attribute's value is in the list of supplied values
attribute_exists	0	the attribute has a value
attribute_not_exists	0	the attribute does not have a value
attribute_type	1	the attribute's value is of the supplied type
begins_with	1	the attribute's value begins with the supplied value
contains	1	the attribute's value is a string that contains the supplied substring or a set that contains the supplied element

See the official documentation for more info.

Projections

Projections can be added to a get, batch get or query request to control what attributes are returned. This saves bandwidth on the request.

For example...

const tableSchema = {
    tableName: "Transactions",
    partitionKeyField: "customerId",
    partitionKeyType: "string",
    sortKeyField: "transactionId",
    sortKeyType: "string",
};

async function getTransactionDates(customerId) {
    const queryRequest = dynameh.requestBuilder.buildQueryInput(tableSchema, customerId);
    const projectedQueryRequest = dynameh.requestBuilder.addProjection(tableSchema, queryRequest, ["date"]);
    // Note that addProjection() does not change the original object.
    // queryRequest != projectedQueryRequest
    
    const queryResponse = await dynamodb.query(projectedQueryRequest).promise();
    const transactions = dynameh.responseUnwrapper.unwrapQueryOutput(queryResponse);
    return transactions.map(t => t.date);
}

getTransactionDates("BusinessFactory");

Date Serialization

Date serialization can be configured by setting dateSerializationFunction on your TableSchema. It's a function that takes in a Date and returns a string or number. By default Dates are serialized as ISO-8601 strings.

For example...

const tableSchema = {
    tableName: "MyTable",
    partitionKeyField: "id",
    partitionKeyType: "number",
    dateSerializationFunction: date => date.toISOString()   // same as default
};