Hi,
The current automatic schema registration may cause a loss of precision. The 
link below is an example: 
https://issues.apache.org/jira/projects/IOTDB/issues/IOTDB-886?filter=allopenissues
 . Currently, The schema of the numeric type will be registered as float 
automatically. If the num is integer and greater than 2^24, it may lose 
precision. 
I’m currently considering if this infer value is greater than 2^24, use double 
as the automatic schema. The pr is like: 
https://github.com/apache/iotdb/pull/1796.
This plan is currently under discussion. Looking forward to your opinions.


Best,
Wangminhao Gou

Reply via email to