fix: decimal stat rounding overflow #2975
Conversation
gruuya
requested review from
wjones127,
roeap,
rtyler,
hntd187 and
ion-elgreco
as code owners
|
ACTION NEEDED delta-rs follows the Conventional Commits specification for release automation. The PR title and description are used as the merge commit message. Please update your PR title and description to match the specification. |
Signed-off-by: Marko Grujic <[email protected]>
gruuya
force-pushed
the
fix-decimal-stat-overflow
branch
from
2bee4af
to
fbc8626
Compare
gruuya
changed the title
| if (val.trunc() as i128).to_string().len() > (precision - scale) as usize { | ||
| // For normal values with integer parts that get rounded to a number beyond | ||
| // the precision - scale range take the next smaller (by magnitude) value | ||
| val = f64::from_bits(val.to_bits() - 1); |
There was a problem hiding this comment.
Not sure how performant this is, but the result won't be consistent/parse-able without it.
The actual "round-down" could be done with next_down once it's stable/delta-rs uses nightly.
There was a problem hiding this comment.
@rtyler would you be ok with having delta-rs use nightly features?
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #2975 +/- ##
==========================================
+ Coverage 72.26% 72.28% +0.01%
==========================================
Files 128 128
Lines 40329 40363 +34
Branches 40329 40363 +34
==========================================
+ Hits 29143 29175 +32
+ Misses 9334 9332 -2
- Partials 1852 1856 +4 ☔ View full report in Codecov by Sentry. |
Description
Try to detect the situation where floating point conversion rounds the integer value beyond the valid Decimal(scale, precision) range, and rollback to next smaller value towards 0.0.
Related Issue(s)
Closes #2974
Documentation
None