Elasticsearch and YAML

2016-08-25 elasticsearch

Most examples in Elasticsearch documentation are using JSON to represent documents, requests and responses. Writing JSON is not that hard but you sometimes become Raiders of the Lost Curly Brace.

Raiders of the Lost Ark

Do you know that you can replace JSON by YAML?

In responses

You can add format=yaml in the query params:

curl -XGET http://localhost:9200/logstash-*/_search?format=yaml

---
took: 13
timed_out: false
_shards:
  total: 15
  successful: 15
  failed: 0
hits:
  total: 2382
  max_score: 1.0
  hits:
  - _index: "logstash-2016.06.22"
    _type: "test"
    _id: "AVV36B3vd6qInGJw3x0r"
    _score: 1.0
    _source:
      '@timestamp': "2016-06-22T11:39:07.431Z"
  - _index: "logstash-2016.06.22"
    _type: "test"
    _id: "AVV36Cfrd6qInGJw3x0u"
    _score: 1.0
    _source:
      '@timestamp': "2016-06-22T11:38:42.000Z"

You can also use the Content-Type: application/yaml HTTP header:

curl -XGET -H "Content-Type: application/yaml" http://localhost:9200/_cluster/health

---
cluster_name: "elasticsearch"
status: "yellow"
timed_out: false
number_of_nodes: 1
number_of_data_nodes: 1
...

While using YAML format for responses you don’t have to use the pretty=true query param, as YAML is naturally indented and human readable.

In requests

Just start your request body with ---:

curl -XPOST http://localhost:9200/_search -d '---
query:
  match:
    message: "Elasticsearch"
'

This gets pretty handy when dealing with mappings and settings:

movie.mapping.yaml
---
properties:
  title:
    type: string
  tags:
    type: string
    index: not_analyzed
  year:
    type: date
    format: year
curl -XPUT http://localhost:9200/movies/_mappings/movie --data-binary @movie.mapping.yaml

Limits

Some requests like _bulk request may not accept YAML.

Sense (Kibana Console) and other Elasticsearch UIs do not handle YAML properly.

Other posts