Index vs Type in Elasticsearch: Key Differences and Usage

In Elasticsearch, a index is a collection of documents that share similar characteristics and is the main container for data storage. A type was a way to logically separate different kinds of documents within an index, but it is now deprecated and removed in recent versions, so index is the primary way to organize data.

Quick Comparison

Here is a quick side-by-side comparison of index and type in Elasticsearch.

Factor	Index	Type
Definition	Main container for documents	Logical category inside an index (deprecated)
Purpose	Stores and organizes data	Separates document kinds within an index
Uniqueness	Unique name per cluster	Unique name per index
Status	Core concept, always used	Deprecated since ES 6.x, removed in ES 7.0
Mapping	Defines fields for all documents	Defined per type (legacy)
Usage Today	Use multiple indexes for separation	Avoid using types; use one mapping per index

Key Differences

An index in Elasticsearch is like a database in a traditional system. It holds all the documents and their data. Each index has its own settings and mappings that define how data is stored and searched.

A type was used to group different document categories inside a single index, similar to tables in a database. However, this caused confusion and technical issues, so Elasticsearch deprecated types starting from version 6 and completely removed them in version 7.

Now, you should create separate indexes for different data categories instead of using types. This simplifies data management and avoids conflicts in field definitions.

Code Comparison

Creating an index with a type and adding a document (legacy approach):

json

PUT /my_index
{
  "mappings": {
    "my_type": {
      "properties": {
        "name": { "type": "text" },
        "age": { "type": "integer" }
      }
    }
  }
}

PUT /my_index/my_type/1
{
  "name": "Alice",
  "age": 30
}

Output

{
  "acknowledged": true,
  "shards_acknowledged": true,
  "index": "my_index"
}

{
  "_index": "my_index",
  "_type": "my_type",
  "_id": "1",
  "_version": 1,
  "result": "created"
}

Index Equivalent

Creating an index without types and adding a document (modern approach):

json

PUT /my_index
{
  "mappings": {
    "properties": {
      "name": { "type": "text" },
      "age": { "type": "integer" }
    }
  }
}

PUT /my_index/_doc/1
{
  "name": "Alice",
  "age": 30
}

Output

{
  "acknowledged": true,
  "shards_acknowledged": true,
  "index": "my_index"
}

{
  "_index": "my_index",
  "_type": "_doc",
  "_id": "1",
  "_version": 1,
  "result": "created"
}

When to Use Which

Choose index as your main data container in Elasticsearch for all current and future projects. Avoid using types because they are deprecated and removed in recent versions. If you need to separate data logically, create multiple indexes instead of relying on types. This approach is simpler, more robust, and fully supported.

Key Takeaways

An index is the main way to organize data in Elasticsearch.

Types are deprecated and removed; avoid using them.

Use separate indexes to separate different data categories.

Modern Elasticsearch uses a single mapping per index without types.

Always prefer the latest Elasticsearch syntax for compatibility.