How to and efficient way to combine standard tokenizer with autocomplete (type ahead) functionality

Trying to use the autocomplete functionality which I have in place using
the following mappings

analysis : filter : placename_ngram : max_gram=15, min_gram = 2, type =
edge_ngram
analyzer :index: filter : lowercase, placename_ngram, tokenizer : keyword
placename_search : filter: lowercase: tokenizer keyword

This works great for type ahead but when I'm trying to find a value like
"contains in" it doesn't return the record.

Such as
If I'm doing a text query on "Lake".

I will only get

Lake...
Lake Wood,

But will not get
Smithtown Lake

I have the field setup as multi-field and can do wildcard to find the
values but not sure if this is efficient.

I believe I can use NGRAM but that seems like alot of overhead considering
I only need index terms by whitespace (or by word). Not every permetation.

Any thoughts?

When I change the tokenizer on both to "standard"....It will then find
these records...but my autocomplete gets messed up and brings back
Smithtown Lake when typing Lak..... (which in this case I don't want).

Thanks for your help

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.