I may be missing this in the documentation, so forgive me.
I want a template that specifies a handful of fields so unwanted fields aren't indexed, in order to save disk space. However, I don't want to explicitly define types for those fields; I just want the default types.
there is not a way to do is, because it is still unsafe. Imagine you had this possibility in combination with dailiy indices. On day 1 the first document contains a 1 as a number (possibly resulting in a short), and on day 2 the first document contains a 3458345643 (an int) and on day 3 it will be abcdef ending up in a string. You would again end up with an arbitrary mapping, just within the limitation of known fields.
If I was responsible for that cluster, I would just go with a solid preconfigured mapping and then decide what to do with unknown fields: either ignoring them or throwing an exception on indexation.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.