I'm searching in two different modes using Thinking Sphinx:
- Full search on a single model for normal search functionality
- Full search across all models for autocomplete dropdown functionality
For the sake of this question, let's say I have a Person and a Country model.
When performing regular searches, I want to fetch all people who's name of country name matches the search string. To achieve this, I have added an index on the countries name in the Person index. All well so far.
When searching to populate my autocomplete dropdown, I want to show all countries and all people matching my search string. Here the problem shows up. When doing an Application-Wide search, I now get:
- all countries whose name match my search string
- all doctors whose name match my search string, and unfortunately...
- all doctors who belongs to a country that matches the search string.
The last part makes for some really confusing autocomplete results for the user. Is there any simple way for me to avoid this by using built-in functionality, for example like having two indices on the Person model, and choose which one to use for each kind of search?