CODEWITHSUNDEEP
solrsearch-engine
Nathan Friend
Nathan Friend

Reputation: 12834

Multiple/split filter chains for a single Solr field

Is it possible to configure a Solr field to have multiple/split Filter chains? For example, could I create a Filter chain that looks like this?:

        StandardTokenizer
                ↓
         LowerCaseFilter
          ↙           ↘
  SynonymnFilter   PhoneticFilter    
        ↓
   NGramFilter

I've done quite a bit of searching and haven't found any examples of setting up a Filter chain this way.

EDIT

The main reason I would like to do this in the context of a single field (as opposed to indexing the data twice using a copy field) is for highlighting. If I use a copy field to run a different filter chain on the same data, my highlighting results come back like this:

"highlighting":{
  "1234": {
    "firstName_phonetic":["<hl>John</hl>"],
    "firstName_ngram":["<hl>John</hl>"]
  }
}

This makes consuming the results a bit more difficult as the consuming app needs to choose which highlighted field to display. Ideally, I'd get something back like this:

"highlighting":{
  "1234": {
    "firstName":["<hl>John</hl>"]
  }
}

Upvotes: 0

Views: 212

Answers (2)

Anand
Anand

Reputation: 81

The output of one filter goes to the next filter in the chain. There is no support for "branching".

You'll be implementing this method if you want a custom filter.

org.apache.lucene.analysis.TokenStream.incrementToken()

This method returns a boolean. You can see an example here.

Another possible solution is to use a copyField to combine firstName_phonetic and firstName_ngram into firstName and use hl.fl=firstName

Upvotes: 1

Persimmonium
Persimmonium

Reputation: 15789

It sounds reasonable just to use copyField and index two different fields, each with its own chain.

But, I guess nothing prevents you from implementing your own MyCustomFilter that does what you need, and you plug that one into your conf.

Upvotes: 1

Related Questions