Does ferret support near real time searching? I want to index documents
and search them as quickly as possible. Currently I am indexing, then
persisting the data and retrieving the results from the disk. I am kind
of able to simulate near real time behavior in ferret - I have my index
as global instance. The indexer first indexes in the RAM and then
persists the data onto disk. In the Searching part I am using multiple
readers - one to search in the RAM and one to search on the disk. I am
then combining these results.
The issue is that multireader is taking very long time to fetch results
from RAM and disk and then merge those results. Is there any way
near-real time indexing can be implemented in ferret?