Index word bigram with count

I would need to index text fields in this way:
given a string "what's the weather like today?" I would like to index a document with all the word bigrams and their counts (also for single words) as fields:

"what's" with count 1
"the" with count 1
"what's the" with count 1
"the weather" with count 1

Then, I need to be able to query for a word (or a bigram) and get its count over all documents.
How should I do?
Thank you!

a shingle filter together with the term suggester might be sufficient for this. The question is why do you need the raw term frequencies?

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