I see... Didn't see it in the docs though. Is it there?
Anyways, I understand doc_values access is way faster than _source, that most use cases deal with values already in doc_values, and some features require hard limits for safety's sake.
What I don't understand is the call for blocking a feature because it is not as performant as another feature. Painless works inline as is, and it is very useful for ad_hoc querying. It is stupid to store a groovy query for a quick question and unsafe to allow it to run inline. I needed a quick answer from a field that is not indexed because I don't need it at query time for my application.
Why don't let the user worry about the consequences? In a SQL database you can easily blow it by issuing a simple SELECT, but you can issue it anyways.
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.