Add decimal support to VARIANT casting - #23858

Merged

rapids-bot[bot] merged 25 commits into

NVIDIA:mainfrom

vuule:variant-decimal-cast

Sep 3, 2026

Commits on Aug 27, 2026

Add decimal support to VARIANT casting
vuule
committed
64be3f0
Address review: int128 and sliced coverage, share cast preamble, benchmark decimals
vuule
committed
1370071
Trim verbose comments in variant decimal cast
vuule
committed
b93a33f
Remove redundant decimal cast tests, extend cast matrix with decimal target
vuule
committed
44c8749
Cover the int64 decimal representation in tests and benchmarks
vuule
committed
df1bcbc
Use cudf::is_fixed_point directly instead of an alias that narrows nothing
vuule
committed
305e6d7

Commits on Aug 29, 2026

Merge remote-tracking branch 'origin/main' into variant-decimal-cast
vuule
committed
79e292e

Commits on Sep 1, 2026

Merge remote-tracking branch 'origin/main' into variant-decimal-cast
vuule
committed
554e2a3
Name the expected scale in the decimal cast tests
vuule
committed
8327c7c
Give expected_scale the scale_type so expectations pass it through
vuule
committed
405e29d
Tidy up the decimal cast tests
vuule
committed
c8c1660
Divide once when rescaling a decimal, in 64 bits where possible
vuule
committed
3fde395
Load the unscaled payload only for a known decimal width
vuule
committed
03e70e0
Merge branch 'main' into variant-decimal-cast
vuule
authored
e818b78

Commits on Sep 2, 2026

comment update

vuule
and
mhaseeb123
authored
b15b724
constexpr1

vuule
and
mhaseeb123
authored
370bd42
constexpr2

vuule
and
mhaseeb123
authored
3e5e445
constexpr3

vuule
and
mhaseeb123
authored
d53483a
Answer the two decided cases of a decimal scale-up up front
vuule
committed
6f348af
Merge branch 'variant-decimal-cast' of https://github.com/vuule/cudf into variant-decimal-cast
vuule
committed
0766f3f
Fix clang-format spacing after the constexpr suggestions
vuule
committed
ba7a2a2
paranoid overflow protection
vuule
committed
3fbdfe4
Merge branch 'main' into variant-decimal-cast
vuule
authored
e8e95d2
Derive the decimal overflow bounds from the representation types
vuule
committed
e5b6dc6
Merge branch 'main' into variant-decimal-cast
vuule
authored
80ea5dc