Synonym indexing VS Payload indexing (and queries)


(Laurent Jakubina) #1

Hello everyone,

I'm interested into adding meta-information "on" tokens/phrases (like tagging metrics info as "measures" or adding concept synonyms to written -- in the initial doc -- concepts). I just read about Synonym indexing (https://www.elastic.co/blog/multitoken-synonyms-and-graph-queries-in-elasticsearch) and Payload indexing (https://lucidworks.com/2009/08/05/getting-started-with-payloads/) and...I'm not sure which one could be the best match to my use cases ?

Synonym indexing seems to be well covered into the Elasticsearch documentation. "Code" samples and blogs cover the available functionalities. It does not seem the case for Payload indexing ; only https://www.elastic.co/guide/en/elasticsearch/reference/6.2/analysis-delimited-payload-tokenfilter.html mention the payloads existence into the docs. But you can find some questions and answers on the web, sometimes.

Synonym indexing seems easy to implement but we do not seem to have all the tools to manage the added contents (for instance, modifying the underlying model -- vector space model -- according to the new content). And it seems the opposite for payloads... To that, I'm not sure that I can add anything I want into the byte arrays of payloads.

Any information, ideas, ... ? Could be nice to have the differences/similarities/advantages/etc between the 2 functionality !

Thanks !


(Laurent Jakubina) #2

No idea or remarks ?

Thanks


(system) #3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.