If lucene is used to build a document database, can it run independently without a mongo-like storage

  mongodb, question


The requirement is similar to this one: full-text search (of course, full-text Chinese) can be carried out as it is.
The problem now is that lucene can also keep the original field contents when indexing.
So comprehensive
* performance (no more than 200 /s queries),
* update (hope data can be reflected immediately after update, but retrieval can be delayed),
* data volume (less than ten million, each data will not be m)
* Extensibility (multi-source, format and field of each source are different)
, do you need to save the original data in the database and lucene only has one primary key, or is it better to restore the original data directly from lucene
I have done similar projects before. I saved the original content in mongo, but I didn’t think much about it.

It is suggested to refer to the implementation of elasticsearch.