is this link a hacky way of doing it though?

pk using email#bobby.tables@gmail.com

It's just different from traditional DBs

My general advice is dynamo is an excellent tool to pair with serverless, but it's not something you can incrementally learn. There's a fairly high upfront cost to learning the patterns and re-using knowledge from relational DBs won't work. If you're new to serverless, it might be a lot to learn serverless and also learn Dynamo at the same time. Would suggest using RDS serverless (which we have a new construct for) or planetscale

so on insert. you add your row data fields: {     subscribeId: sst.TableFieldType.STRING,     emailAddress: sst.TableFieldType.STRING,     confirmed: sst.TableFieldType.BOOL,     timestamp: sst.TableFieldType.STRING,    }, then you add another fields: {     subscribeId: sst.TableFieldType.STRING,    }, where subscribeId = emailAddress#Example@123.com ?

yes...but I wouldn't have the table structure you have to begin with since I'd opt for single table design

And I usually use a library like this: https://github.com/sensedeep/dynamodb-onetable Or this one which I haven't tried yet: https://github.com/tywalch/electrodb

I need to go away and hit dyamodb and learn the onetable pattern 🙂

thanks alot

lol...

Preface My DynamoDB story begins with an abundance of unfounded confidence, slowly beaten out of me by the club of experience.

don't worry, some of us made it all the way to production on unfounded confidence

To my readers born after 1995, a "phone book" was a real-life physical book containing the names, addresses, and phone numbers of everyone in a particular location. The phone book was in alphabetical order by last name. Imagine that! This was before fancy things like Google and Facebook rendered phone books obsolete for anything other than kindling.

@Derek Kershner it would solve my problem but understanding how to leverage dynamodb correctly was the more valuable lesson in this thread, specifically the single table design principle. Which comes with learning to create ERD diagrams and understanding access patterns to design an application with an efficient model

its quite cool because initially I want to capture someone's email address and that's all I will know about them, but later on if they register in my application I want to link that user email address to that person who registered. So Im starting my single table with only an email address, but then once they register, I have more data about them, once they complete their profile, I again have more info, and once they go though a shopping cart I again have more information. at the same time they might never register, or they might register but never sign up to the newsletter lol

and then theres the fact that people can have multiple email addresses 😄

i could use a complex type for the multiple email addresses actually

instead of email -> example@123.com i could do email -> [{example@123.com,mysecondemail@123.com}]

do you use hooks to keep in sync with cognito too? say if someone changes something in cognito? or do you stick to immutable data?

what I dont quite understand yet is if you need to create the whole schema in advance into the table or if you can just add to the single table

like if i deployed a table with id, tmp Can i add to the table with something not define? like create into table id, tmp, beta

where id = id

do you have default fields set too?

in the initial bit

Error: No fields defined for the "SingleTable_V2" Table

this.table = new sst.Table(this, "SingleTable_V2", {        fields: {       //   userId: sst.TableFieldType.STRING,       //   email: sst.TableFieldType.STRING,       },       primaryIndex: { partitionKey: "PK", sortKey: "SK"},     });

would it be this.table = new sst.Table(this, "SingleTable_V2", {        fields: {          PK: sst.TableFieldType.STRING,          SK: sst.TableFieldType.STRING,       },       primaryIndex: { partitionKey: "PK", sortKey: "SK"},     }); ???

oooh think i got it!

so i did do this.table = new sst.Table(this, "SingleTable_V2", {        fields: {          pk: sst.TableFieldType.STRING,          sk: sst.TableFieldType.STRING,       },       primaryIndex: { partitionKey: "pk", sortKey: "sk"},     });

then had some jiggling about

created a route

routes: {         "POST /test": {           function: {             srcPath: "services/test/",             handler: "create.main",             environment: { TABLE_NAME: props.table.tableName },             permissions: [props.table],           },         },

then i created the js

import handler from "../../util/handler"; import {Table} from 'dynamodb-onetable' import AWS from "aws-sdk"; const client = new AWS.DynamoDB.DocumentClient(); export const main = handler(async (event) => {     const MySchema = {       format: 'onetable:1.1.0',       version: '0.0.1',       indexes: {           primary: { hash: 'pk', sort: 'sk' },           gs1:     { hash: 'gs1pk', sort: 'gs1sk', follow: true },           ls1:     { sort: 'id', local: true },       },       models: {           Account: {               pk:          { type: String, value: 'account:${name}' },               sk:          { type: String, value: 'account:' },               //id:          { type: String, generate: 'ulid', validate: /^[0-9A-F]{32}$/i },               id:          { type: String, generate: 'ulid' },               name:        { type: String, required: true },               status:      { type: String, default: 'active' },               zip:         { type: String },           },           User: {               pk:          { type: String, value: 'account:${accountName}' },               //sk:          { type: String, value: 'user:${email}', validate: EmailRegExp },               sk:          { type: String, value: 'user:${email}' },               id:          { type: String, required: true },               accountName: { type: String, required: true },               email:       { type: String, required: true },               firstName:   { type: String, required: true },               lastName:    { type: String, required: true },               username:    { type: String, required: true },               role:        { type: String, enum: ['user', 'admin'], required: true, default: 'user' },               balance:     { type: Number, default: 0 },               gs1pk:       { type: String, value: 'user-email:${email}' },               gs1sk:       { type: String, value: 'user:' },           }       },       params: {           'isoDates': true,           'timestamps': true,       },   }   const table = new Table({     client: client,     name: 'kldev-OxyDjinn-SingleTable_V2',     schema: MySchema, })   //CREATE   const Account = table.getModel('Account')   let account = await Account.create({     id: '2ab73092-82bc-47e6-8271-481457eca6e3',     name: 'Acme Airplanes',   })   return account; });

(working on crud testing)

and that did work nicely!

(ignore the post -d i havnt wired that bit up yet)

curl -X POST -H 'Content-Type: application/json' -d '{"emailAddress":"Example@123.com"}' https://blah.execute-api.us-east-1.amazonaws.com/test {"id":"2ab73092-82bc-47e6-8271-481457eca6e3","name":"Acme Airplanes","status":"active","created":"2022-02-12T220639.254Z","updated":"2022-02-12T220639.254Z"}

woot! the main gotcha for me was a bunch of permission errors because of the name of the table name: 'kldev-OxyDjinn-SingleTable_V2', but once i figured that it started working

thanks so much

what would be the best practise for this

// Scan is not normally advised -- scans entire table let users = await User.scan({}) console.log('FOUND users', users) for (let user of users) { await User.remove({id: user.id}) }

say if i was an admin on an admin console and i wanted to get a list of users for some reason

This is why you need to define your data access patterns up front so you can design the data structures and key overloading to get what you need

Sometimes this is via a GSI where you do things like reversing PK and SK. There lots of good examples in The DynamoDB Book

have you ever used onetable and needed to get the cancellation reason from a transact? im struggling to get the reason and google isnt helping me one bit 😞

let user = await table.transact('write', transaction).catch((error) => {       let errorMessage = 'Could not create user'       console.log(error)       console.log(transaction.TransactItems)       return {         error: errorMessage       }     })

OneTableError: OneTable execute failed "transactWrite" for "_Generic. Transaction cancelled, please refer cancellation reasons for specific reasons [ConditionalCheckFailed, None]

i cant seem to get the cancellation reasons

lol thanks Ross i will give that a go

let user = await table.transact('write', transaction).catch((error) => {     let errorMessage = 'Could not create user'     if(error.context.err.code === "TransactionCanceledException")     {       //we have to split the err.message to get the transactional failures. a null is OK but anything else is not.       var errors = error.context.err.message.split('[')[1].split(']')[0].split(',')       if(errors[0] === 'ConditionalCheckFailed') {         errorMessage = 'User with this username already exists.'       } else if (errors[1] === 'ConditionalCheckFailed'){         errorMessage = 'User with this email already exists.'       }       console.log(error.context.err.message)     }         return {       error: errorMessage     }   })

if(errors[0].Code === 'ConditionalCheckFailed') {         errorMessage = 'User with this username already exists.'       } else if (errors[1].Code === 'ConditionalCheckFailed'){         errorMessage = 'User with this email already exists.'       }