Queries against ‘string’ fields in rt-index

Hi, I’m new to manticore, looks great, very promising - thanks to all the developers for their efforts!

Can you give me an overview on the performance of queries against ‘string’ fields in a rt-index? I am looking up a ‘key’ field with an exact value and i am not sure if this is as optimized as for example a query against a ‘unique index field’ in a mysql-db. Currently i have only ~100000 documents in my first testbed and all works fine but there could be millions too - how to handle such case, maybe better use a different index(-type)?

Many Thanks

String attributes are not “fast” as there is no index on them yet. Add the key as an indexed text (full-text search) and perform matching. At 100K docs you might not see a big difference between the two, but at 1M the query times should start to be different.

Thanks for the advice! Regarding “no index on them yet” - are there any plans on this?

we have prototypes on attribute indexes however do not have exact plans on developing that feature for now