# Using Elasticsearch for indexing

**URL:** <https://community.bloomreach.com/t/using-elasticsearch-for-indexing/398>\
**Category:** Experience Manager (PaaS/OnPrem)\
**Created:** [September 12, 2018, 2:45pm UTC](https://community.bloomreach.com/t/using-elasticsearch-for-indexing/398 "2018-09-12T14:45:22Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![mriegler](https://yyz1.discourse-cdn.com/flex027/user_avatar/community.bloomreach.com/mriegler/32/178_2.png) [@mriegler](https://community.bloomreach.com/u/mriegler)\
**Post date:** [September 12, 2018, 2:45pm UTC](https://community.bloomreach.com/t/using-elasticsearch-for-indexing/398/1 "2018-09-12T14:45:22Z")

</div>

The documentation recommends using Elasticsearch as the search provider, but has no documentation regarding it compared to Solr.

Therefore I have some conceptual questions. My biggest issue is the concept of a schema/type/mapping in Elasticsearch related to DocumentBeans in Hippo. Would I need to create a separate index with its own type for each document type in Hippo? If I don’t, I’d end up with a very sparse (a.k.a. most documents only have a few fields, but the entire schema has all possible fields) index.

On the other hand, if I created an index for each document type (news, events, etc.) I’m looking at a lot of manual mapping back and forth, creating separate indices and having to search multiple indices at a time.

Does anyone have any experience with managing this problem? Having multiple document types that you need to search?

---

<div class="post-metadata">

**Author:** ![jasper.floor](https://yyz1.discourse-cdn.com/flex027/user_avatar/community.bloomreach.com/jasper.floor/32/28_2.png) [@jasper.floor](https://community.bloomreach.com/u/jasper.floor)\
**Post date:** [September 14, 2018, 1:30pm UTC](https://community.bloomreach.com/t/using-elasticsearch-for-indexing/398/2 "2018-09-14T13:30:25Z")

</div>

You already have a Lucene index available, please make sure that cannot cover your needs before using an external engine. That being said, I don’t think you would need an index per document type, but I am not the expert you are looking for. I think this question has more to do with ES than with BloomReach, so maybe ask on an ES board?