Querying Recurring Events In A Fauna FQL Database - database

Question

I'm having some trouble querying recurring events using Fauna DB / FQL. I'm storing the events like so:
{
id: 1,
userId: 1,
title: "A Very Cool Title",
description: "A Basic Description",
date: {
day: 23,
month: 11,
year: 2022,
hour: 0,
minute: 0
},
frequency: {
minutes: 1,
hours: 1,
days: 1,
weeks: 1,
months: 1,
years: 1
}
}
The date is the original date of this event, and the frequency is how often it occurs. The frequency can be as low as 1 minute. I would like to be able to lookup all events that fall between a start and end date for a specific userId.
My first instinct is to just add the frequency to the date and check if the sum of that falls within the queried start and end date, repeating that until it exceeds the end date. However, since my frequency can be as low as 1 minute, I would have to repeat that a crazy number of times and it just seems inefficient.
This could be a problem with querying but it also may simply be a problem with storing the events differently, I'm not sure.

Answer 1

I would like to be able to lookup all events that fall between a start and end date for a specific userId.
Except for the starting events, all of your events are virtual and not amenable to "lookup". Instead, you'd have to calculate events.
There is no do { ... } while condition in FQL, and there are transaction limits that would prohibit processing many events.
You should definitely use the Timestamp type instead of an object to record dates. Then you can use the various Time() or Date() functions to perform the necessary time calculations.
Your model includes multiple levels of frequency, minutes, hours, etc. The problem is notably harder to solve if an event can recur every 1 year, 1 month, 1 week, 1 day, 1 hour, and 1 minute.
It becomes easier when you limit the expression of the frequency to an amount and a unit. For example:
frequency: { amount: 3, unit: "days" }
With that model, you could determine the number of events with TimeDiff(start_date, end_date, unit). For example, if an event started on October 17, the number of daily events between then and now would be:
> TimeDiff(Time("2022-10-17T00:00:00Z"), Now(), "days")
37
If you need to produce entries for each virtual event, then you'd bump into another limitation of FQL: there is no for loop. There is ForEach() or Map(), but you have to already have a set/array to iterate on.
So, you would have to combine a couple of tricks.
Once you have the number of virtual events, you can use Repeat() to generate a string based on a template that includes a delimiter:
> Repeat("a ", 37)
'a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a '
With that string, you can use RTrim() to remove the trailing space.
The you can use the recipe for creating a SplitString function. The you can call the function:
> Call("SplitString", "a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a", " ")
[
'a', 'a', 'a', 'a', 'a', 'a', 'a',
'a', 'a', 'a', 'a', 'a', 'a', 'a',
'a', 'a', 'a', 'a', 'a', 'a', 'a',
'a', 'a', 'a', 'a', 'a', 'a', 'a',
'a', 'a', 'a', 'a', 'a', 'a', 'a',
'a', 'a'
]
Now you have an array that can be used for the scope of the events, and iterate with Reduce() to start at the starting date, use TimeAdd() to add the frequency interval once for each entry in the array, and compute the virtual event's timestamp.
Without knowing the structure of any index you might be using for your event documents, I can't provide a query that demonstrates the complete effort, but these tips should help with the mechanics of the computation.