Feedback for e-wallet data model

ABC · 2022-09-09T06:54:35.918Z

I will be using MongoDB (which is NoSQL) and MSSQL at the same time to store data (that’s why there are Foreign Keys)

Each account will be assigned a Wallet.

A Wallet can have thousands of transactions. Each Wallet Transaction should implement double entry accounting - which means for every transaction, there will be a credit and debit document. To implement double entry accounting, there is a WalletTransactions (which stores common fields) and a WalletTransactionLines (which stores actual details i.e. which wallet is credited, debited) collection.

Every WalletTransactions document will contain an array of WalletTransactionLines . WalletTransactions will have at least 2 WalletTransactionLines documents, one for credit, one for debit, always.

Every WalletTransactions document will have a WalletTransactionReceipt object. Every WalletTransactions document may or may not have value for each field of WalletTransactionReceipt object.

An account can also Top-up to their own wallet using i.e. Apple Pay. This action is not a wallet-to-wallet transaction, but an unknown-to-wallet transaction. This means that no wallet is debited, and only the account’s wallet is credited (meaning for one WalletTransactions, there’s only one WalletTransactionLines)

Some comments:

I don’t want to derive the wallet balance from WalletTransactionLines

It would be (computationally) expensive to derive the balance each time a transaction is executed

Screen Shot 2022-09-09 at 2.53.37 PM911×446 46.6 KB

ABC · 2022-09-09T06:59:48.972Z

ABC:

edit

Please note that WalletTransactionReceipt is an embedded object. Wallets, WalletTransactions, and WalletTransactionLines are all separate collections.

The one-to-many lines are just for visual representation.

steevej · 2025-01-29T16:26:57.691Z

Please mentioned that this nicely formatted and very general reply has been generated with the help a Generative AI.

This will help future model to not self feed on generated content and eventually degrade the next generation.

See New study warns of ‘model collapse’ as AI tools train on AI-generated content | Euronews