If I want to make a messaging app like WhatsApp would bubble be good for it I’m just asking based on the workload units wouldn’t that be used up quickly ?
I always thought that bubble was smart enough to not do repeat searches. As you mentioned in the video, we should use custom states to hold all the data when no filters are selected and then filter it further using the values in the custom state, but wouldn’t bubble have downloaded all that data already to the page? Meaning it wouldn’t have to fetch more data if all filters are cleared again, it would just use the retrieved data it already downloaded. I did try testing this out, and the app metrics only shows WU’s for the initial download of data. I would play around with many filters constantly for 10 minutes straight, and only the first minute shows any kind of WU consumption. I hope Bubble’s not hiding it from me😅 What are your thoughts?
Retriving all the datat in a table in database (and store it in a custom state then do filtering) can cost a lot when there is a lot of data, so I think there is no effective way to do searches exept data structure. Complex filtering requires complex data structure and one search with no :filtered
Wouldn’t this prevent realtime viewing of new objects that have been added to the database? Could certainly have the state refresh every 30 seconds or so, but which is more labor intensive?
Correct, if you're dependent on a state, then you'll need to manage updating it as well. Some search scenarios don't need to worry about this as much. If you need keep up with constant real-time changes, then you'll want to focus on optimizing the constraints in a direct search as much as possible. Searching isn't necessarily bad! You just don't want to over-search, so keep an eye out for opportunities to consolidate throughout the app.
This video doesn't cover daisy chaining filters with workflows, but that's a great approach for when you have more advanced filtering needs involving many inputs, multiple data types, and conditional logic. The same optimization concepts in this video would still apply though.
