A query with
inner_hits is only returning the
_source segment from within all the
hits, and the
inner_hits are being ignored. The ScrollReader only has logic to extract
id fields, see here - elasticsearch-hadoop/ScrollReader.java at 0db74fdf51ef96f43b8648e6d1116a4a39469ee1 · elastic/elasticsearch-hadoop · GitHub
Can this be expanded to support inner_hits segment as well?