I have the honor to program a plant filter search.
Means you can search plants meeting certain conditions (attributes).
But I need some advice for counting number of occurring attributes.
Over simplified records:
- id: 1; attributes: [color: [red, green], size: 10]; - id: 2; attributes: [color: [red], light: [moderate];
The attributes are varying!
The filter search functionality should provide a counter for each attribute:
color: red (2), green (1) size: 10 (1) light: moderate (1)
Now if you’ll select
size will be (0),
Sadly PostgreSQL is getting to its limit with this requirement.
Counting attributes is too slow.
So I’m in search of an alternative approach. → Advice needed!
- It concerns about 2,5 k records
- Attributes of all records will have a size of “just” about 10 - 15 MB
- The first most common filters will reduce the record amount just by about 20%
- Update of records / attributes won’t be very frequent
Should I just cache a list of all records + attributes in for example ETS?
And than do the counting / filtering with some
Or are there any other good approaches I’m currently blind for?