Hello,
I am unsure if solr is the right solution for a problem
that we have, of if it is better to stick with a relational
database (and if it should be done in solr how to implement it).
The explanation is a bit lengthy, but please
bear with me.
The problem:
Sort results of a vendor search for a product according to price
including delivery costs.
The data:
The store itself is a marketplace, and each product can be
supplied by different vendors. The vendors can define delivery
costs for different price ranges.
E.g:
_price from_| price to_| delivery cost |
| 0 | 49 | 10 |
vendor --| 50 | 99 | 5 |
| 100 | max | 0 |
So, for product with a price of 55, I want the result to be 60.
Additional requirements:
- The product price is also calculated, based on properties
of the vendor.
- There is also a pickup option, and there should be no
duplicate results.
- Different shipping costs for different countries.
Progress so far:
My idea is to store each range as a subdocument for a vendor, but
I don't know how to construct a query for that. So far I have
managed to implement a simpler version that gives the right result for
each country using dynamic fields, but this uses only a free delivery
above x approach and that is not what we want.
I have looked into the Block Join Query parser, but as far as I can
tell this does not allow to construct a function query with inputs
from parent and child documents.
Why solr:
- sort and limit result according to geolocation.
- we will deploy solr anyhow in this project, for a classic
full text search.
As said above, I'm not really sure if this is a good application
for solr, but the geolocation features are quite handy. And the
query is not really fast in a relational db either.
Any input is greatly appreciated.
Regards,
Ewald
