[
https://issues.jboss.org/browse/TEIID-5692?page=com.atlassian.jira.plugin...
]
Steven Hawkins resolved TEIID-5692.
-----------------------------------
Resolution: Done
Generally speaking the old behavior was attempting to preserve the approximate type scale
- but that is not necessary and at worst creates values where the scale is much larger
than expected.
org.teiid.preserveApproximateScale was added as a system/environment property should this
need to be toggled to the old behavior. It doesn't seem necessary at this point to
make this a documented property.
Scaling float/double to bigdecimal/numeric is not always necessary
------------------------------------------------------------------
Key: TEIID-5692
URL:
https://issues.jboss.org/browse/TEIID-5692
Project: Teiid
Issue Type: Quality Risk
Components: Query Engine
Reporter: Steven Hawkins
Assignee: Steven Hawkins
Priority: Major
Fix For: 12.2
There is logic in the FloatingNumberToBigDecimalTransform to set the scale of the result
bigdecimal to something representative of the scale of a float/double. However the way
that java represents precision/scale of small and large value - for example 0.0d becomes
0E-15 after the transformation - implies more precision or scale than expected.
--
This message was sent by Atlassian Jira
(v7.12.1#712002)