Skip to content

GH-46557: [C++][R] Check overflow in float to int casts that allow truncation #52469

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
advitrocks9 wants to merge 6 commits into
base: main
Choose a base branch
from
Open

GH-46557: [C++][R] Check overflow in float to int casts that allow truncation #52469

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
d490bd6
GH-46557: [C++][R] Check overflow in float to int casts that allow tr…
advitrocks9 Oct 9, 2026
d255ed4
Keep safe cast error messages unchanged
advitrocks9 Oct 9, 2026
a7a29ee
Disable float-cast-overflow sanitizer in half float to int cast
advitrocks9 Oct 9, 2026
95a2688
Decode half float values in cast error messages
advitrocks9 Oct 11, 2026
f2a1baa
GH-46557: [C++] Check the range before converting when truncation is …
advitrocks9 Oct 22, 2026
f65ceaa
GH-46557: [C++] Check the range after the conversion
advitrocks9 Oct 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions cpp/src/arrow/compute/kernels/scalar_cast_internal.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,7 @@ struct CastPrimitive<HalfFloatType, InType, enable_if_integer<InType>> {
// Cast half float to int
template <typename OutType>
struct CastPrimitive<OutType, HalfFloatType, enable_if_integer<OutType>> {
ARROW_DISABLE_UBSAN("float-cast-overflow")
static void Exec(const ArraySpan& arr, ArraySpan* out) {
using OutT = typename OutType::c_type;
const uint16_t* in_values = arr.GetValues<uint16_t>(1);
Expand Down
111 changes: 81 additions & 30 deletions cpp/src/arrow/compute/kernels/scalar_cast_numeric.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,8 @@

// Implementation of casting to integer, floating point, or decimal types

#include <cmath>

#include "arrow/array/builder_primitive.h"
#include "arrow/compute/kernels/common_internal.h"
#include "arrow/compute/kernels/scalar_cast_internal.h"
Expand Down Expand Up @@ -72,6 +74,11 @@ struct WasTruncated {
static bool CheckMaybeNull(OutT out_val, InT in_val, bool is_valid) {
return is_valid && static_cast<InT>(out_val) != in_val;
}

static Status Error(InT in_val, const DataType& out_type) {
return Status::Invalid("Float value ", in_val, " was truncated converting to ",
out_type);
}
};

// Half float to int
Expand All @@ -85,18 +92,60 @@ struct WasTruncated<HalfFloatType, OutType> {
static bool CheckMaybeNull(OutT out_val, uint16_t in_val, bool is_valid) {
return is_valid && static_cast<float>(out_val) != Float16::FromBits(in_val).ToFloat();
}

static Status Error(uint16_t in_val, const DataType& out_type) {
return WasTruncated<FloatType, OutType>::Error(Float16::FromBits(in_val).ToFloat(),
out_type);
}
};

// InType is a floating point type we are planning to cast to integer
template <typename InType, typename OutType, typename InT = typename InType::c_type,
typename OutT = typename OutType::c_type>
ARROW_DISABLE_UBSAN("float-cast-overflow")
Status CheckFloatTruncation(const ArraySpan& input, const ArraySpan& output) {
auto GetErrorMessage = [&](InT val) {
return Status::Invalid("Float value ", val, " was truncated converting to ",
*output.type);
};
struct WasOutOfRange {
static_assert(std::is_integral_v<OutT>);

static constexpr auto kMin = static_cast<InT>(std::numeric_limits<OutT>::min());

@pitrou pitrou Oct 22, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These seems to assume that OutT is integral, can we add a static_assert about that?

static constexpr auto kMaxPlusOne =
static_cast<InT>(std::numeric_limits<OutT>::max() / 2 + 1) * 2;

static bool Check(OutT, InT in_val) {
return !(std::trunc(in_val) >= kMin) | !(in_val < kMaxPlusOne);
}

static bool CheckMaybeNull(OutT out_val, InT in_val, bool is_valid) {
return is_valid && Check(out_val, in_val);
}

static Status Error(InT in_val, const DataType& out_type) {
return Status::Invalid("Float value ", in_val, " out of range converting to ",
out_type);
}
};

template <typename OutType>
struct WasOutOfRange<HalfFloatType, OutType> {
using OutT = typename OutType::c_type;
static bool Check(OutT out_val, uint16_t in_val) {
return WasOutOfRange<FloatType, OutType>::Check(out_val,
Float16::FromBits(in_val).ToFloat());
}

static bool CheckMaybeNull(OutT out_val, uint16_t in_val, bool is_valid) {
return is_valid && Check(out_val, in_val);
}

static Status Error(uint16_t in_val, const DataType& out_type) {
return WasOutOfRange<FloatType, OutType>::Error(Float16::FromBits(in_val).ToFloat(),
out_type);
}
};

// InType is a floating point type we are planning to cast to integer
template <template <typename...> class Checker, typename InType, typename OutType,
typename InT = typename InType::c_type,
typename OutT = typename OutType::c_type>
ARROW_DISABLE_UBSAN("float-cast-overflow")
Status CheckFloatToIntValues(const ArraySpan& input, const ArraySpan& output) {
const InT* in_data = input.GetValues<InT>(1);
const OutT* out_data = output.GetValues<OutT>(1);

Expand All @@ -110,29 +159,28 @@ Status CheckFloatTruncation(const ArraySpan& input, const ArraySpan& output) {
if (block.popcount == block.length) {
// Fast path: branchless
for (int64_t i = 0; i < block.length; ++i) {
block_out_of_bounds |=
WasTruncated<InType, OutType>::Check(out_data[i], in_data[i]);
block_out_of_bounds |= Checker<InType, OutType>::Check(out_data[i], in_data[i]);
}
} else if (block.popcount > 0) {
// Indices have nulls, must only boundscheck non-null values
for (int64_t i = 0; i < block.length; ++i) {
block_out_of_bounds |= WasTruncated<InType, OutType>::CheckMaybeNull(
block_out_of_bounds |= Checker<InType, OutType>::CheckMaybeNull(
out_data[i], in_data[i], bit_util::GetBit(bitmap, offset_position + i));
}
}
if (ARROW_PREDICT_FALSE(block_out_of_bounds)) {
if (input.GetNullCount() > 0) {
for (int64_t i = 0; i < block.length; ++i) {
if (WasTruncated<InType, OutType>::CheckMaybeNull(
if (Checker<InType, OutType>::CheckMaybeNull(
out_data[i], in_data[i],
bit_util::GetBit(bitmap, offset_position + i))) {
return GetErrorMessage(in_data[i]);
return Checker<InType, OutType>::Error(in_data[i], *output.type);
}
}
} else {
for (int64_t i = 0; i < block.length; ++i) {
if (WasTruncated<InType, OutType>::Check(out_data[i], in_data[i])) {
return GetErrorMessage(in_data[i]);
if (Checker<InType, OutType>::Check(out_data[i], in_data[i])) {
return Checker<InType, OutType>::Error(in_data[i], *output.type);
}
}
}
Expand All @@ -145,41 +193,42 @@ Status CheckFloatTruncation(const ArraySpan& input, const ArraySpan& output) {
return Status::OK();
}

template <typename InType>
Status CheckFloatToIntTruncationImpl(const ArraySpan& input, const ArraySpan& output) {
template <template <typename...> class Checker, typename InType>
Status CheckFloatToIntImpl(const ArraySpan& input, const ArraySpan& output) {
switch (output.type->id()) {
case Type::INT8:
return CheckFloatTruncation<InType, Int8Type>(input, output);
return CheckFloatToIntValues<Checker, InType, Int8Type>(input, output);
case Type::INT16:
return CheckFloatTruncation<InType, Int16Type>(input, output);
return CheckFloatToIntValues<Checker, InType, Int16Type>(input, output);
case Type::INT32:
return CheckFloatTruncation<InType, Int32Type>(input, output);
return CheckFloatToIntValues<Checker, InType, Int32Type>(input, output);
case Type::INT64:
return CheckFloatTruncation<InType, Int64Type>(input, output);
return CheckFloatToIntValues<Checker, InType, Int64Type>(input, output);
case Type::UINT8:
return CheckFloatTruncation<InType, UInt8Type>(input, output);
return CheckFloatToIntValues<Checker, InType, UInt8Type>(input, output);
case Type::UINT16:
return CheckFloatTruncation<InType, UInt16Type>(input, output);
return CheckFloatToIntValues<Checker, InType, UInt16Type>(input, output);
case Type::UINT32:
return CheckFloatTruncation<InType, UInt32Type>(input, output);
return CheckFloatToIntValues<Checker, InType, UInt32Type>(input, output);
case Type::UINT64:
return CheckFloatTruncation<InType, UInt64Type>(input, output);
return CheckFloatToIntValues<Checker, InType, UInt64Type>(input, output);
default:
break;
}
DCHECK(false);
return Status::OK();
}

Status CheckFloatToIntTruncation(const ExecValue& input, const ExecResult& output) {
template <template <typename...> class Checker>
Status CheckFloatToInt(const ExecValue& input, const ExecResult& output) {
switch (input.type()->id()) {
case Type::FLOAT:
return CheckFloatToIntTruncationImpl<FloatType>(input.array, *output.array_span());
return CheckFloatToIntImpl<Checker, FloatType>(input.array, *output.array_span());
case Type::DOUBLE:
return CheckFloatToIntTruncationImpl<DoubleType>(input.array, *output.array_span());
return CheckFloatToIntImpl<Checker, DoubleType>(input.array, *output.array_span());
case Type::HALF_FLOAT:
return CheckFloatToIntTruncationImpl<HalfFloatType>(input.array,
*output.array_span());
return CheckFloatToIntImpl<Checker, HalfFloatType>(input.array,
*output.array_span());
default:
break;
}
Expand All @@ -192,7 +241,9 @@ Status CastFloatingToInteger(KernelContext* ctx, const ExecSpan& batch, ExecResu
CastNumberToNumberUnsafe(batch[0].type()->id(), out->type()->id(), batch[0].array,
out->array_span_mutable());
if (!options.allow_float_truncate) {
RETURN_NOT_OK(CheckFloatToIntTruncation(batch[0], *out));
RETURN_NOT_OK(CheckFloatToInt<WasTruncated>(batch[0], *out));
} else if (!options.allow_int_overflow) {
RETURN_NOT_OK(CheckFloatToInt<WasOutOfRange>(batch[0], *out));
}
return Status::OK();
}
Expand Down
60 changes: 60 additions & 0 deletions cpp/src/arrow/compute/kernels/scalar_cast_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -398,6 +398,66 @@ TEST(Cast, FloatingToInt) {
}
}

TEST(Cast, FloatingToIntOverflow) {
for (auto from : {float16(), float32(), float64()}) {
for (bool allow_float_truncate : {false, true}) {
auto opts = CastOptions::Safe(int8());
opts.allow_float_truncate = allow_float_truncate;
for (auto json : {"[128.0]", "[-129.0]", "[NaN]", "[Inf]", "[-Inf]"}) {

@pitrou pitrou Oct 22, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we also test 127.5 and -128.5 when truncation is disabled?

CheckCastFails(ArrayFromJSON(from, json), opts);
}
CheckCast(ArrayFromJSON(from, "[127.0, null, -128.0]"),
ArrayFromJSON(int8(), "[127, null, -128]"), opts);
CheckCast(TweakValidityBit(ArrayFromJSON(from, "[1000.0, 1.0]"), 0, false),
ArrayFromJSON(int8(), "[null, 1]"), opts);

opts.to_type = uint8();
CheckCastFails(ArrayFromJSON(from, "[256.0]"), opts);
CheckCastFails(ArrayFromJSON(from, "[-1.0]"), opts);
Comment on lines +415 to +416

@pitrou pitrou Oct 22, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What about 255.0 and -0.5?

CheckCast(ArrayFromJSON(from, "[255.0, null, 0.0]"),
ArrayFromJSON(uint8(), "[255, null, 0]"), opts);
}

auto opts = CastOptions::Safe(int8());
CheckCastFails(ArrayFromJSON(from, "[127.5]"), opts);
CheckCastFails(ArrayFromJSON(from, "[-128.5]"), opts);
opts.allow_float_truncate = true;
CheckCast(ArrayFromJSON(from, "[127.5, -128.5]"),
ArrayFromJSON(int8(), "[127, -128]"), opts);

opts = CastOptions::Safe(uint8());
CheckCastFails(ArrayFromJSON(from, "[255.5]"), opts);
CheckCastFails(ArrayFromJSON(from, "[-0.5]"), opts);
opts.allow_float_truncate = true;
CheckCast(ArrayFromJSON(from, "[255.5, -0.5]"), ArrayFromJSON(uint8(), "[255, 0]"),
opts);
}

for (auto from : {float32(), float64()}) {
auto opts = CastOptions::Safe(int64());
opts.allow_float_truncate = true;
for (auto json : {"[9223372036854775807]", "[9223372036854775807.5]",
"[9223372036854775808.0]"}) {
CheckCastFails(ArrayFromJSON(from, json), opts);
}
CheckCast(ArrayFromJSON(from, "[-9223372036854775808.0, -9223372036854775808.5]"),
ArrayFromJSON(int64(), "[-9223372036854775808, -9223372036854775808]"),
opts);
opts.to_type = uint64();
CheckCastFails(ArrayFromJSON(from, "[18446744073709551616.0]"), opts);
}

auto opts = CastOptions::Safe(int8());
EXPECT_RAISES_WITH_MESSAGE_THAT(
Invalid,
::testing::HasSubstr("Float value 1.500000 was truncated converting to int8"),
Cast(ArrayFromJSON(float16(), "[1.5]"), opts));
opts.allow_float_truncate = true;
EXPECT_RAISES_WITH_MESSAGE_THAT(
Invalid, ::testing::HasSubstr("Float value inf out of range converting to int8"),
Cast(ArrayFromJSON(float16(), "[Inf]"), opts));
}

TEST(Cast, FloatingToFloating) {
for (auto from : {float16(), float32(), float64()}) {
for (auto to : {float16(), float32(), float64()}) {
Expand Down
2 changes: 1 addition & 1 deletion r/R/expression.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -211,7 +211,7 @@ Expression$op <- function(FUN, ..., args = list(...)) {
"if_else",
Expression$op("==", args[[2]], 0L),
Scalar$create(NA_integer_, out_type),
cast(out, out_type, allow_float_truncate = TRUE)
cast(out, out_type, allow_float_truncate = TRUE, allow_int_overflow = TRUE)
)
}
return(out)
Expand Down
Loading