In the documentation for 'search_after', it says that inconsistent results may be returned if a refresh occurs when not using a PIT. I have a few questions:
Can duplicate results be returned?
Will it return all the documents that were present before the refresh (excluding any documents that were deleted in the refresh)?
I'm using '_doc' as the sort order, and I'm using the 'sort' value of the last hit in every search_after request. Will I eventually get all documents that were present when the first request was made, without duplicates, given that they weren't deleted in a refresh?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.