I'm trying to optimize nested term aggregations for an autocomplete use case. More info on the query and mapping here: https://stackoverflow.com/questions/65613335/improve-performance-of-a-nested-term-aggregation.
Based on https://github.com/elastic/elasticsearch/issues/22912#issuecomment-277189389 it seems like it is recommended to create a separate index for autocomplete on search terms which may be duplicated across documents. Is that still the case?