Apache Solr模式配置

Question

So I am pretty new to Apache Solr and have a situation I do now know how to handle. I am from an OO programming background so first let me explain the object relationships:

Take an object called Movie that has two text fields, title and description. A movie can be associated with tags by a user. These tags are particular to the user, and are not visible to other users.

So an example Movie could have something like this:

"Movie Title", "Description of the Movie" User1Tags: "tag1", "tag2" User2Tags: "action", "somethingElse"

I need to design a schema/solr query so that when user1 is searching for movies, if they type "action", the movie above will not show up. This is because user2 has associated "action" with "Movie Title", not user1.

Things I have considered:

1) Filter queries - these do not seem to work as once the index per movie is built, I do not see how to avoid having all the user tags be tied to the movie's index.

2) A separate core for movie to tag associations and just doing two queries per search. I know I can do it this way, but making another core seems excessive to me.

Are there other options I am missing? Or is there a way to implement 1? Or is the simplest option just option 2 and that's how people who know what they are doing with Solr would do it?

Answer 1

How many users?

If not many, then you can have dynamic fields tag_user1, tag_user2 and modify the eDismax field list to match or not match against it, eg by using field name alias.

The other option is to prefix the values with the userid. So tags field would have: user1_tag1, user1_tag2, user2_action, user2_somethingElse. Then, you need a custom filter in the query chain that will prefix your search tokens with the user of the request and so only prefixed values would match.