# Autocomplete on large fields

**URL:** <https://discuss.elastic.co/t/autocomplete-on-large-fields/8855>\
**Category:** Elasticsearch\
**Created:** [August 27, 2012, 1:56am UTC](https://discuss.elastic.co/t/autocomplete-on-large-fields/8855 "2012-08-27T01:56:24Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Keith\_Webster](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/keith_webster/32/1532_2.png) [@Keith\_Webster](https://discuss.elastic.co/u/Keith_Webster)\
**Post date:** [August 27, 2012, 1:56am UTC](https://discuss.elastic.co/t/autocomplete-on-large-fields/8855/1 "2012-08-27T01:56:24Z")

</div>

Hey guys,

There are several different posts in relation to autocomplete so I do  
apologise for creating yet another.

I've read through all the autocomplete posts on the forum and the general  
consensus is not to use wildcards or a prefix filter since they are so  
memory intensive and slow but instead use an edge n gram filter.

Since I'm building an autocomplete that will search approx 10 fields, 2 of  
which are large description fields (hundreds of words per field and  
potentially different versions for different languages) is edge n gram  
still the way to go? Obviously the index size will be much larger but will  
the query still be efficient if I use a max gram of 14? How will memory  
usage be impacted by all this or will it be impacted at all?

Any thoughts/advice would be great.

Thanks

--

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [July 6, 2017, 3:14am UTC](https://discuss.elastic.co/t/autocomplete-on-large-fields/8855/2 "2017-07-06T03:14:59Z")

</div>