A while ago, I had to add a discount system for a web app. The discount period is a week and the start can differ from a user to another. Later on, we want to filter those users who are on a discount and those who are not.
The task seemed simple and quickly decided on using the following approach.
- Create a discount
datefield in the user document on the DB (MongoDb) - After the week has passed, replace the discount date field to a
booleanset tofalse.
The reason behind replacing a date field to boolean is to save up myself some lines of code and avoid dates comparison which can be tricky to compare sometimes, especially with multiple timezones. So I started implementing, finished, tested then deployed to production.
A few days later, I found out getting more data from the discount system can be useful. Imagine all the information about how well/bad the discount affects the users number over time, just by knowing when the users' discounts were created and ended. The problem is I only have the start date until the discount is finished, then neither start nor end dates are available anymore. How sad..
So I changed quickly to a better design, scratch off number 2 and add the
start and end dates. Let the code be longer, but now I got
more reliable and specific data.
Lesson learnt
- Having dates about events in the database gives you more insight for your data
- Troubleshooting on production level can be tricky without enough information
- Give enough time for designing your Database fields and how data flows with some use cases.
Top comments (0)