Using Ngrams for building a autocomplete suggester

I have 2 fields in my index, Title and Description.

Title obviously a short phrase (max 100 characters) and description is a long description (up to 5000 characters).

I have built a autocomplete by building ngrams on Title fields. So any word in the title field will be used to build ngrams. The autocomplete suggester works fine, but I am not sure if it would be a good idea to include description fields in autocomplete suggester?

Pros: description contains more words, so if I build ngrams on description field, I would get suggestions on a lot more words

Cons: Description is quite long and I am not sure if building so many ngrams is going to slow down Elasticsearch?

Is there any best practice performance wise when choosing fields that we are using to build ngrams? Is it recommended to avoid long fields (such as description in this example) in building ngrams?

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