Idea for the API:

• Require starting Hibernate Search, which can be done offline, see and

• Add the feature to the schema manager

E.g. we’d add methods such as this one to the schema manager:

And calling such methods would result in the creation of files in the given directory, with a predictable directory structure:

Of course the content of the export would be backend-specific; for Lucene, it would just do nothing (or maybe create the directory structure but leave a file for each index whose content says “the Lucene backend does not support exporting the schema”).

Later we might add more options (filtering, …) and features (export curl scripts, …), but let’s wait a bit before working on that: exporting json should be enough for most people.

Thank you for this feedback !

Hey , I knew you had requested this, but when I implemented the feature I couldn’t find your message… I just stumbled upon it by chance, so here’s a (very very late) answer: you can now manage schemas manually in Hibernate Search 6 ( ) and there is also a flag to drop-and-create the schema when mass indexing ( , use .purge(false).dropAndCreateSchemaOnStart(true))

I hope this helps…

Hello,

Related to this this CLI feature, public API to allow index dropping and schema creation (index by index) may be an appreciated feature to handle manually drop, create schema then reindex that may have better performance than a delete from match-all query (big performance hit for ES 2.4.5, much better performance for ES 5.x).