使用 Lambda 对 DynamoDB 数据扫描进行排序答案

【问题标题】：Sort DynamoDB data scan using Lambda使用 Lambda 对 DynamoDB 数据扫描进行排序
【发布时间】：2020-02-06 06:08:41
【问题描述】：

我在 DynamoDB 中创建了一个表，主键是一个存储为字符串的数字。当我使用 DynamoClient 的 GetItem 时，项目的顺序是随机的。

这是我的 Lambda 函数；

'use strict';
const AWS = require('aws-sdk');

exports.handler = async (event, context) => {
    // Bring in the documentClient library
    const documentClient = new AWS.DynamoDB.DocumentClient(); 

    // instanialize 
    let responseBody = "";
    let statusCode = 0;

    // extract variables that we need to use in db
    // const { country, visits } = JSON.parse(event.body);

    const params = {
        TableName: "age_distribution"
    };

    try {
        const data = await documentClient.scan(params).promise(); 
        // add data into the db
        responseBody = JSON.stringify(data.Items); 
        // hold the response in a string type
        statusCode = 200;
    } catch (err) {
        responseBody = `Unable to get record: ${err}`;
        statusCode = 403;
    }

    const response = {
        statusCode: statusCode,
        headers: {
            "Content-Type" : " application/json",
            "access-control-allow-origin" : "*"
        },
        body: responseBody 
        //send back the responseBody
    };

    // return the response
    return response;
};

这是示例响应

[{"country":"63","visits":3},{"country":"92","visits":4},{"country":"64","visits":3},{"country":"90","visits":5},{"country":"79","visits":11},{"country":"66","visits":10},{"country":"80","visits":20},{"country":"62","visits":4},{"country":"87","visits":6},{"country":"75","visits":22},{"country":"76","visits":16},{"country":"97","visits":1},{"country":"84","visits":13},{"country":"60","visits":2},{"country":"61","visits":4},{"country":"89","visits":7},{"country":"94","visits":1},{"country":"98","visits":1},{"country":"82","visits":17},{"country":"77","visits":16},{"country":"91","visits":4},{"country":"81","visits":18},{"country":"88","visits":10},{"country":"96","visits":1},{"country":"71","visits":18},{"country":"95","visits":1},{"country":"78","visits":21},{"country":"83","visits":12},{"country":"68","visits":14},{"country":"65","visits":6},{"country":"70","visits":12},{"country":"85","visits":9},{"country":"69","visits":13},{"country":"74","visits":12},{"country":"93","visits":3},{"country":"86","visits":10},{"country":"67","visits":6},{"country":"72","visits":11},{"country":"73","visits":26}]

我想对项目进行排序，从 60 开始到 98 结束

【问题讨论】：

标签： javascript amazon-dynamodb

【解决方案1】：

DynamoDb 中的排序是针对索引的排序键值进行的

From the docs:

查询结果始终按排序键值排序。如果排序键的数据类型为Number，则以数字顺序返回结果；否则，结果将按 UTF-8 字节的顺序返回。默认情况下，排序顺序是升序。要反转顺序，请将 ScanIndexForward 参数设置为 false

所以您可能想要做的是创建一个 GSI（全局二级索引），其中包含对您有意义的排序键（范围）我经常在此索引定义中使用时间戳：

    GlobalSecondaryIndexes:
        - IndexName: productKey-timestamp-index
        KeySchema:
          - AttributeName: productKey
            KeyType: HASH
          - AttributeName: timestamp
            KeyType: RANGE

然后，当我查询此索引时，我会根据时间戳获取数据升序系列。

如果我希望它降序，那么我使用以下查询：

 let params = {
        TableName: ORDERS_TABLE,
        Limit: 10,
        IndexName: 'productKey-timestamp-index',
        Key: {
            "productKey": productKey
        },
        KeyConditionExpression: 'productKey= :productKey',
        ExpressionAttributeValues: {
            ':productKey': productKey
        },
        ScanIndexForward: false
    };

    db.query(params).then(console.log)

注意：设置ScanIndexForward: false 会通过排序键反转顺序，即在这种情况下为timestamp

【讨论】：