You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In scenarios with multiple nodes, shards and replicas >= 1, search results of hybrid query with sorting do have duplicates or inconsistent records (same field has different values when returned in document source and sorting field sections).
I was able to replicate this with the cluster with 2 nodes, 5 primary and 1 replica shards. Search request has both sort and search_after.
Reproduced in latest 2.x (2.19 snapshot) and main.
example of response. Wrong data is in the last 8-th document, value of trendingScore is 0.0 when it's part of the source, and 3.0 when the same field is part of the sort section
What is the bug?
In scenarios with multiple nodes, shards and replicas >= 1, search results of hybrid query with sorting do have duplicates or inconsistent records (same field has different values when returned in document source and sorting field sections).
I was able to replicate this with the cluster with 2 nodes, 5 primary and 1 replica shards. Search request has both
sort
andsearch_after
.Reproduced in latest 2.x (2.19 snapshot) and main.
Here is the search request:
example of response. Wrong data is in the last 8-th document, value of
trendingScore
is 0.0 when it's part of the source, and 3.0 when the same field is part of thesort
sectionThe text was updated successfully, but these errors were encountered: