Skip to content

Commit

Permalink
Folder: type, vector
Browse files Browse the repository at this point in the history 
relative pr:

Support more data types for read filter oap-project#139
Fix cast double to decimal oap-project#179
Fix casting from string to decimal oap-project#281
Support cast decimal to int oap-project#177
Fix null on overflow and multiply as spark precision and support cast varchar to decimal oap-project#169
Disable tokenizing the path by dot oap-project#109
Serialize and deserialize RowVector oap-project#250
Support datetime pattern in spark oap-project#94
  • Loading branch information
@zhejiangxiaomai
zhejiangxiaomai committed May 31, 2023
1 parent f5e31fa commit c20fa1d
Show file tree
Hide file tree
Showing 22 changed files with 1,466 additions and 59 deletions.
154 changes: 139 additions & 15 deletions velox/type/Conversions.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,12 +21,17 @@
#include <string>
#include <type_traits>
#include "velox/common/base/Exceptions.h"
#include "velox/type/DecimalUtil.h"
#include "velox/type/TimestampConversion.h"
#include "velox/type/Type.h"

namespace facebook::velox::util {

template <TypeKind KIND, typename = void, bool TRUNCATE = false>
template <
TypeKind KIND,
typename = void,
bool TRUNCATE = false,
bool ALLOW_DECIMAL = false>
struct Converter {
template <typename T>
// nullOutput API requires that the user has already set nullOutput to
Expand All @@ -37,12 +42,26 @@ struct Converter {
VELOX_UNSUPPORTED(
"Conversion to {} is not supported", TypeTraits<KIND>::name);
}

template <typename T>
static typename TypeTraits<KIND>::NativeType
cast(T val, bool& nullOutput, const TypePtr& toType) {
VELOX_UNSUPPORTED(
"Conversion of {} to {} is not supported",
CppToType<T>::name,
TypeTraits<KIND>::name);
}
};

template <>
struct Converter<TypeKind::BOOLEAN> {
using T = bool;

template <typename From>
static T cast(const From& v, bool& nullOutput, const TypePtr& toType) {
VELOX_NYI();
}

template <typename From>
static T cast(const From& v, bool& nullOutput) {
return folly::to<T>(v);
Expand All @@ -69,24 +88,61 @@ struct Converter<TypeKind::BOOLEAN> {
}
};

template <TypeKind KIND, bool TRUNCATE>
template <TypeKind KIND, bool TRUNCATE, bool ALLOW_DECIMAL>
struct Converter<
KIND,
std::enable_if_t<
KIND == TypeKind::BOOLEAN || KIND == TypeKind::TINYINT ||
KIND == TypeKind::SMALLINT || KIND == TypeKind::INTEGER ||
KIND == TypeKind::BIGINT || KIND == TypeKind::HUGEINT,
void>,
TRUNCATE> {
TRUNCATE,
ALLOW_DECIMAL> {
using T = typename TypeTraits<KIND>::NativeType;

template <typename From>
static T cast(const From& v, bool& nullOutput, const TypePtr& toType) {
VELOX_NYI();
}

// from long decimal cast to some type
static T cast(const int128_t& d, bool& nullOutput, const TypePtr& fromType) {
const auto& decimalType = fromType->asLongDecimal();
auto scale0Decimal = DecimalUtil::rescaleWithRoundUp<int128_t, int128_t>(
d,
decimalType.precision(),
decimalType.scale(),
decimalType.precision(),
0,
false,
false);
return cast(scale0Decimal.value(), nullOutput);
}

// from short decimal cast to some type
static T cast(const int64_t& d, bool& nullOutput, const TypePtr& fromType) {
const auto& decimalType = fromType->asShortDecimal();
auto scale0Decimal = DecimalUtil::rescaleWithRoundUp<int64_t, int64_t>(
d,
decimalType.precision(),
decimalType.scale(),
decimalType.precision(),
0,
false,
false);
return cast(scale0Decimal.value(), nullOutput);
}

template <typename From>
static T cast(const From& v, bool& nullOutput) {
VELOX_UNSUPPORTED(
"Conversion to {} is not supported", TypeTraits<KIND>::name);
}

static T convertStringToInt(const folly::StringPiece& v, bool& nullOutput) {
static T convertStringToInt(
const folly::StringPiece& v,
const bool allowDecimal,
bool& nullOutput) {
// Handling boolean target case fist because it is in this scope
if constexpr (std::is_same_v<T, bool>) {
return folly::to<T>(v);
Expand All @@ -110,6 +166,10 @@ struct Converter<
}
if (negative) {
for (; index < len; index++) {
// Allow decimal and ignore the fractional part.
if (v[index] == '.' && allowDecimal) {
break;
}
if (!std::isdigit(v[index])) {
return -1;
}
Expand All @@ -121,6 +181,9 @@ struct Converter<
}
} else {
for (; index < len; index++) {
if (v[index] == '.' && allowDecimal) {
break;
}
if (!std::isdigit(v[index])) {
return -1;
}
Expand All @@ -140,7 +203,7 @@ struct Converter<
static T cast(const folly::StringPiece& v, bool& nullOutput) {
try {
if constexpr (TRUNCATE) {
return convertStringToInt(v, nullOutput);
return convertStringToInt(v, ALLOW_DECIMAL, nullOutput);
} else {
return folly::to<T>(v);
}
Expand All @@ -152,7 +215,8 @@ struct Converter<
static T cast(const StringView& v, bool& nullOutput) {
try {
if constexpr (TRUNCATE) {
return convertStringToInt(folly::StringPiece(v), nullOutput);
return convertStringToInt(
folly::StringPiece(v), ALLOW_DECIMAL, nullOutput);
} else {
return folly::to<T>(folly::StringPiece(v));
}
Expand All @@ -164,7 +228,7 @@ struct Converter<
static T cast(const std::string& v, bool& nullOutput) {
try {
if constexpr (TRUNCATE) {
return convertStringToInt(v, nullOutput);
return convertStringToInt(v, ALLOW_DECIMAL, nullOutput);
} else {
return folly::to<T>(v);
}
Expand Down Expand Up @@ -221,7 +285,9 @@ struct Converter<
if (v > LimitType::maxLimit()) {
return LimitType::max();
}
if (v < LimitType::minLimit()) {
// bool type's min is 0, but spark expects true for casting negative float
// data.
if (!std::is_same_v<T, bool> && v < LimitType::minLimit()) {
return LimitType::min();
}
return LimitType::cast(v);
Expand All @@ -241,7 +307,9 @@ struct Converter<
if (v > LimitType::maxLimit()) {
return LimitType::max();
}
if (v < LimitType::minLimit()) {
// bool type's min is 0, but spark expects true for casting negative float
// data.
if (!std::is_same_v<T, bool> && v < LimitType::minLimit()) {
return LimitType::min();
}
return LimitType::cast(v);
Expand Down Expand Up @@ -284,15 +352,39 @@ struct Converter<
return folly::to<T>(v);
}
}

static T cast(const int128_t& v, bool& nullOutput) {
if constexpr (TRUNCATE) {
return T(v);
} else {
return static_cast<T>(v);
}
}
};

template <TypeKind KIND, bool TRUNCATE>
template <TypeKind KIND, bool TRUNCATE, bool ALLOW_DECIMAL>
struct Converter<
KIND,
std::enable_if_t<KIND == TypeKind::REAL || KIND == TypeKind::DOUBLE, void>,
TRUNCATE> {
TRUNCATE,
ALLOW_DECIMAL> {
using T = typename TypeTraits<KIND>::NativeType;

template <typename From>
static T cast(const From& v, bool& nullOutput, const TypePtr& toType) {
VELOX_NYI();
}

static T cast(const int64_t& v, bool& nullOutput, const TypePtr& fromType) {
auto decimalType = fromType->asShortDecimal();
return DecimalUtil::toDoubleValue(v, decimalType.scale());
}

static T cast(const int128_t& v, bool& nullOutput, const TypePtr& fromType) {
auto decimalType = fromType->asLongDecimal();
return DecimalUtil::toDoubleValue(v, decimalType.scale());
}

template <typename From>
static T cast(const From& v, bool& nullOutput) {
try {
Expand Down Expand Up @@ -358,10 +450,31 @@ struct Converter<
VELOX_UNSUPPORTED(
"Conversion of Timestamp to Real or Double is not supported");
}

static T cast(const int128_t& d, bool& nullOutput) {
VELOX_UNSUPPORTED(
"Conversion of int128_t to Real or Double is not supported");
}
};

template <bool TRUNCATE>
struct Converter<TypeKind::VARCHAR, void, TRUNCATE> {
template <bool TRUNCATE, bool ALLOW_DECIMAL>
struct Converter<TypeKind::VARCHAR, void, TRUNCATE, ALLOW_DECIMAL> {
template <typename T>
static std::string
cast(const T& v, bool& nullOutput, const TypePtr& fromType) {
VELOX_NYI();
}

static std::string
cast(const int64_t& v, bool& nullOutput, const TypePtr& fromType) {
return DecimalUtil::toString(v, fromType);
}

static std::string
cast(const int128_t& v, bool& nullOutput, const TypePtr& fromType) {
return DecimalUtil::toString(v, fromType);
}

template <typename T>
static std::string cast(const T& val, bool& nullOutput) {
if constexpr (
Expand Down Expand Up @@ -390,6 +503,11 @@ template <>
struct Converter<TypeKind::TIMESTAMP> {
using T = typename TypeTraits<TypeKind::TIMESTAMP>::NativeType;

template <typename From>
static T cast(const From& v, bool& nullOutput, const TypePtr& toType) {
VELOX_NYI();
}

template <typename From>
static T cast(const From& /* v */, bool& nullOutput) {
VELOX_UNSUPPORTED("Conversion to Timestamp is not supported");
Expand All @@ -415,9 +533,15 @@ struct Converter<TypeKind::TIMESTAMP> {
};

// Allow conversions from string to DATE type.
template <bool TRUNCATE>
struct Converter<TypeKind::DATE, void, TRUNCATE> {
template <bool TRUNCATE, bool ALLOW_DECIMAL>
struct Converter<TypeKind::DATE, void, TRUNCATE, ALLOW_DECIMAL> {
using T = typename TypeTraits<TypeKind::DATE>::NativeType;

template <typename From>
static T cast(const From& v, bool& nullOutput, const TypePtr& toType) {
VELOX_NYI();
}

template <typename From>
static T cast(const From& /* v */, bool& nullOutput) {
VELOX_UNSUPPORTED("Conversion to Date is not supported");
Expand Down
Loading

0 comments on commit c20fa1d

Please sign in to comment.