Basically we have an index serves as a logging system. The logging record doesn't have uniq attribute like an id. So different logging records can be exactly same. But they are NOT technically duplicate records.
With that uniq change, chewy import would remove docs that have exact same body silently.
Expected behavior
Records with same body shouldn't be removed when bulk import.
Actual behavior
Records with same body are removed when bulk import.
Steps to reproduce the problem
Assume we have
class Log
include ActiveModel::Model
include ActiveModel::Serialization
attr_accessor :user_id, :product_id, :action_type
end
class LogIndex < Chewy::Index
index_scope Log
field :user_id, type: "long"
field :product_id, type: "long"
field :action_type, type: "keyword"
end
We found a breaking change when upgrade from 7.2.3 to 7.2.4. It's coming from this particular change https://github.com/toptal/chewy/compare/v7.2.3...v7.2.4#diff-1eea599674257ec549622b98226f87c29e1467eddf042b787451b575f0432807R31
Basically we have an index serves as a logging system. The logging record doesn't have uniq attribute like an id. So different logging records can be exactly same. But they are NOT technically duplicate records. With that
uniq
change, chewy import would remove docs that have exact same body silently.Expected behavior
Records with same body shouldn't be removed when bulk import.
Actual behavior
Records with same body are removed when bulk import.
Steps to reproduce the problem
Assume we have
Then in rails console
It says 2 docs were indexed. But in actual, only 1 doc was successfully indexed.