Token attributes


(Fabiano Nunes) #1

Hello,

I'm coming from "pure" lucene and diving into elasticsearch/solr for
scalability purporses.

Now, I'm trying to store some attributes found in a XML document. More
specifically, token coordinates for future highlighting.
Example: I have a XML with this structure:

elastic
search
community

I want to store the @c attribute from word element (coordinates
left,width,top,height) in each token. So, in the search, I'll retrieve
each
token position for highlighting. There is any good approach to do
this?
In Lucene, I'm using payloads, but I don't know use them in
elasticsearch.


(Shay Banon) #2

This is more complicated to expose in elasticsearch. You will need to build your own analyzer that knows how to parse this and store the relevant data. The problem then is that there isn't a formal hook point to extract it and return it from a search request.
On Tuesday, March 8, 2011 at 3:26 AM, Fabiano Nunes wrote:

Hello,

I'm coming from "pure" lucene and diving into elasticsearch/solr for
scalability purporses.

Now, I'm trying to store some attributes found in a XML document. More
specifically, token coordinates for future highlighting.
Example: I have a XML with this structure:

elastic
search
community

I want to store the @c attribute from word element (coordinates
left,width,top,height) in each token. So, in the search, I'll retrieve
each
token position for highlighting. There is any good approach to do
this?
In Lucene, I'm using payloads, but I don't know use them in
elasticsearch.


(system) #3