I am currently evaluating multiple search/indexing solutions and the one thing I can't find any the mailing list or in any blogs/doc pages is improving leading wildcard query times. Several other implementations of API's on top of Lucene offer some different solutions (sometimes doing a forward and reverse index etc.) to improve query times. Is there something similar implemented in Elastic? Maybe something I missed in the index docs? I checked the settings there for things like storage but it is more about where and FS type settings. Nothing on how to implement anything like what I was looking for?
It looks like I get about a 10 sec penalty if I do a query like:
_all:foo
over if I do the a query like this:
_all:foo* or even
_all:foo
Is this just the cost of doing a leading wildcard or is there a way around taking that performance hit?