On http://www.elasticsearch.org/guide/reference/mapping/ there is a
reference to the schemaless nature of ES and its ability to guess
field types. But how exactly does it guess? Does it try to create
instances of Java types like Date, Integer, Float, Boolean, String,
etc.? Or does it use some heuristic based on field names (there is an
example of "first_name" at the above URL, for example)?
Sematext is hiring Search Engineers -- http://sematext.com/about/jobs.html