Skip to content
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.

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

Pick smallest decimal type with required precision in ORC reader #9775

Merged
merged 19 commits into from
Dec 8, 2021
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
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
24 changes: 24 additions & 0 deletions cpp/include/cudf/io/orc.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -72,6 +72,7 @@ class orc_reader_options {

// Columns that should be read as Decimal128
std::vector<std::string> _decimal128_columns;
bool _enable_decimal128 = true;

friend orc_reader_options_builder;

Expand Down Expand Up @@ -151,6 +152,11 @@ class orc_reader_options {
*/
std::vector<std::string> const& get_decimal128_columns() const { return _decimal128_columns; }

/**
* @brief Whether to use row index to speed-up reading.
*/
bool is_enabled_decimal128() const { return _enable_decimal128; }

// Setters

/**
Expand Down Expand Up @@ -225,6 +231,13 @@ class orc_reader_options {
_decimal_cols_as_float = std::move(val);
}

/**
* @brief Enable/Disable the use of decimal128 type
*
* @param use Boolean value to enable/disable.
*/
void enable_decimal128(bool use) { _enable_decimal128 = use; }

/**
* @brief Set columns that should be read as 128-bit Decimal
*
Expand Down Expand Up @@ -362,6 +375,17 @@ class orc_reader_options_builder {
return *this;
}

/**
* @brief Enable/Disable use of decimal128 type
*
* @param use Boolean value to enable/disable.
*/
orc_reader_options_builder& decimal128(bool use)
{
options.enable_decimal128(use);
return *this;
}

/**
* @brief move orc_reader_options member once it's built.
*/
Expand Down
37 changes: 23 additions & 14 deletions cpp/src/io/orc/reader_impl.cu
Original file line number Diff line number Diff line change
Expand Up @@ -230,24 +230,35 @@ size_t gather_stream_info(const size_t stripe_index,
/**
* @brief Determines cuDF type of an ORC Decimal column.
*/
auto decimal_column_type(const std::vector<std::string>& float64_columns,
const std::vector<std::string>& decimal128_columns,
cudf::io::orc::metadata& metadata,
auto decimal_column_type(std::vector<std::string> const& float64_columns,
std::vector<std::string> const& decimal128_columns,
bool is_decimal128_enabled,
cudf::io::orc::detail::aggregate_orc_metadata const& metadata,
int column_index)
{
auto const& column_path = metadata.column_path(column_index);
if (metadata.get_col_type(column_index).kind != DECIMAL) return type_id::EMPTY;

auto const& column_path = metadata.column_path(0, column_index);
auto is_column_in = [&](const std::vector<std::string>& cols) {
return std::find(cols.cbegin(), cols.cend(), column_path) != cols.end();
};

auto const user_selected_float64 = is_column_in(float64_columns);
auto const user_selected_decimal128 = is_column_in(decimal128_columns);
auto const user_selected_decimal128 = is_decimal128_enabled and is_column_in(decimal128_columns);
CUDF_EXPECTS(not user_selected_float64 or not user_selected_decimal128,
"Both decimal128 and float64 types selected for column " + column_path);

if (user_selected_float64) return type_id::FLOAT64;
if (user_selected_decimal128) return type_id::DECIMAL128;
return type_id::DECIMAL64;

auto const precision = metadata.get_col_type(column_index)
.precision.value_or(cuda::std::numeric_limits<int64_t>::digits10);
if (precision <= cuda::std::numeric_limits<int32_t>::digits10) return type_id::DECIMAL32;
if (precision <= cuda::std::numeric_limits<int64_t>::digits10) return type_id::DECIMAL64;
CUDF_EXPECTS(is_decimal128_enabled,
"Decimal precision too high for decimal64, use `decimal_cols_as_float` or enable "
"decimal128 use");
return type_id::DECIMAL128;
}

} // namespace
Expand Down Expand Up @@ -744,7 +755,7 @@ std::unique_ptr<column> reader::impl::create_empty_column(const size_type orc_co
_use_np_dtypes,
_timestamp_type.id(),
decimal_column_type(
_decimal_cols_as_float, decimal128_columns, _metadata.per_file_metadata[0], orc_col_id));
_decimal_cols_as_float, decimal128_columns, is_decimal128_enabled, _metadata, orc_col_id));
int32_t scale = 0;
std::vector<std::unique_ptr<column>> child_columns;
std::unique_ptr<column> out_col = nullptr;
Expand Down Expand Up @@ -795,7 +806,7 @@ std::unique_ptr<column> reader::impl::create_empty_column(const size_type orc_co
break;

case orc::DECIMAL:
if (type == type_id::DECIMAL64 or type == type_id::DECIMAL128) {
if (type == type_id::DECIMAL32 or type == type_id::DECIMAL64 or type == type_id::DECIMAL128) {
scale = -static_cast<int32_t>(_metadata.get_types()[orc_col_id].scale.value_or(0));
}
out_col = make_empty_column(data_type(type, scale));
Expand Down Expand Up @@ -889,6 +900,7 @@ reader::impl::impl(std::vector<std::unique_ptr<datasource>>&& sources,
// Control decimals conversion
_decimal_cols_as_float = options.get_decimal_cols_as_float();
decimal128_columns = options.get_decimal128_columns();
is_decimal128_enabled = options.is_enabled_decimal128();
}

timezone_table reader::impl::compute_timezone_table(
Expand Down Expand Up @@ -953,13 +965,10 @@ table_with_metadata reader::impl::read(size_type skip_rows,
_use_np_dtypes,
_timestamp_type.id(),
decimal_column_type(
_decimal_cols_as_float, decimal128_columns, _metadata.per_file_metadata[0], col.id));
_decimal_cols_as_float, decimal128_columns, is_decimal128_enabled, _metadata, col.id));
CUDF_EXPECTS(col_type != type_id::EMPTY, "Unknown type");
CUDF_EXPECTS(
(col_type != type_id::DECIMAL64) or (_metadata.get_col_type(col.id).precision <= 18),
"Precision of column " + std::string{_metadata.column_name(0, col.id)} +
" is over 18, use 128-bit Decimal.");
if (col_type == type_id::DECIMAL64 or col_type == type_id::DECIMAL128) {
if (col_type == type_id::DECIMAL32 or col_type == type_id::DECIMAL64 or
col_type == type_id::DECIMAL128) {
// sign of the scale is changed since cuDF follows c++ libraries like CNL
// which uses negative scaling, but liborc and other libraries
// follow positive scaling.
Expand Down
7 changes: 4 additions & 3 deletions cpp/src/io/orc/reader_impl.hpp
Original file line number Diff line number Diff line change
Expand Up @@ -219,12 +219,13 @@ class reader::impl {
cudf::io::orc::detail::aggregate_orc_metadata _metadata;
cudf::io::orc::detail::column_hierarchy selected_columns;

bool _use_index = true;
bool _use_np_dtypes = true;
bool _use_index{true};
bool _use_np_dtypes{true};
std::vector<std::string> _decimal_cols_as_float;
std::vector<std::string> decimal128_columns;
bool is_decimal128_enabled{true};
data_type _timestamp_type{type_id::EMPTY};
reader_column_meta _col_meta;
reader_column_meta _col_meta{};
};

} // namespace orc
Expand Down
14 changes: 8 additions & 6 deletions cpp/src/io/orc/stripe_data.cu
Original file line number Diff line number Diff line change
Expand Up @@ -1066,12 +1066,12 @@ static __device__ int Decode_Decimals(orc_bytestream_s* bs,
return (v / kPow5i[scale]) >> scale;
}
}();
if (dtype_id == type_id::DECIMAL64) {
if (dtype_id == type_id::DECIMAL32) {
vals.i32[t] = scaled_value;
} else if (dtype_id == type_id::DECIMAL64) {
vals.i64[t] = scaled_value;
} else {
{
vals.i128[t] = scaled_value;
}
vals.i128[t] = scaled_value;
}
}
}
Expand Down Expand Up @@ -1708,8 +1708,10 @@ __global__ void __launch_bounds__(block_size)
case DOUBLE:
case LONG: static_cast<uint64_t*>(data_out)[row] = s->vals.u64[t + vals_skipped]; break;
case DECIMAL:
if (s->chunk.dtype_id == type_id::FLOAT64 or
s->chunk.dtype_id == type_id::DECIMAL64) {
if (s->chunk.dtype_id == type_id::DECIMAL32) {
static_cast<uint32_t*>(data_out)[row] = s->vals.u32[t + vals_skipped];
} else if (s->chunk.dtype_id == type_id::FLOAT64 or
s->chunk.dtype_id == type_id::DECIMAL64) {
static_cast<uint64_t*>(data_out)[row] = s->vals.u64[t + vals_skipped];
} else {
// decimal128
Expand Down
23 changes: 9 additions & 14 deletions cpp/tests/io/orc_test.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -387,9 +387,7 @@ TEST_F(OrcWriterTest, MultiColumn)
cudf_io::write_orc(out_opts);

cudf_io::orc_reader_options in_opts =
cudf_io::orc_reader_options::builder(cudf_io::source_info{filepath})
.use_index(false)
.decimal128_columns({"decimal_pos_scale", "decimal_neg_scale"});
cudf_io::orc_reader_options::builder(cudf_io::source_info{filepath}).use_index(false);
auto result = cudf_io::read_orc(in_opts);

CUDF_TEST_EXPECT_TABLES_EQUAL(expected, result.tbl->view());
Expand Down Expand Up @@ -1178,9 +1176,9 @@ TEST_F(OrcWriterTest, Decimal32)
auto data = cudf::detail::make_counting_transform_iterator(0, [&vals](auto i) {
return numeric::decimal32{vals[i], numeric::scale_type{2}};
});
auto mask = cudf::detail::make_counting_transform_iterator(0, [](auto i) { return i % 13 == 0; });
auto mask = cudf::detail::make_counting_transform_iterator(0, [](auto i) { return i % 13; });
column_wrapper<numeric::decimal32> col{data, data + num_rows, mask};
cudf::table_view expected({static_cast<cudf::column_view>(col)});
cudf::table_view expected({col});

auto filepath = temp_env->get_temp_filepath("Decimal32.orc");
cudf_io::orc_writer_options out_opts =
Expand All @@ -1192,12 +1190,7 @@ TEST_F(OrcWriterTest, Decimal32)
cudf_io::orc_reader_options::builder(cudf_io::source_info{filepath});
auto result = cudf_io::read_orc(in_opts);

auto data64 = cudf::detail::make_counting_transform_iterator(0, [&vals](auto i) {
return numeric::decimal64{vals[i], numeric::scale_type{2}};
});
column_wrapper<numeric::decimal64> col64{data64, data64 + num_rows, mask};

CUDF_TEST_EXPECT_COLUMNS_EQUAL(col64, result.tbl->view().column(0));
CUDF_TEST_EXPECT_COLUMNS_EQUAL(col, result.tbl->view().column(0));
}

TEST_F(OrcStatisticsTest, Overflow)
Expand Down Expand Up @@ -1438,7 +1431,7 @@ TEST_F(OrcReaderTest, DecimalOptions)
cudf_io::orc_reader_options::builder(cudf_io::source_info{filepath})
.decimal128_columns({"dec", "fake_name"})
.decimal_cols_as_float({"decc", "fake_name"});
// Should not throw
// Should not throw, even with "fake name" in both options
EXPECT_NO_THROW(cudf_io::read_orc(valid_opts));

cudf_io::orc_reader_options invalid_opts =
Expand Down Expand Up @@ -1493,10 +1486,12 @@ TEST_F(OrcWriterTest, DecimalOptionsNested)
cudf_io::orc_reader_options in_opts =
cudf_io::orc_reader_options::builder(cudf_io::source_info{filepath})
.use_index(false)
.decimal128_columns({"lists.1.dec128"});
.decimal128_columns({"lists.1.dec64"});
auto result = cudf_io::read_orc(in_opts);

CUDF_TEST_EXPECT_TABLES_EQUAL(expected, result.tbl->view());
// Both columns should be read as decimal128
CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(result.tbl->view().column(0).child(1).child(0),
result.tbl->view().column(0).child(1).child(1));
}

CUDF_TEST_PROGRAM_MAIN()
6 changes: 2 additions & 4 deletions python/cudf/cudf/_lib/cpp/io/orc.pxd
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ cdef extern from "cudf/io/orc.hpp" \
void enable_use_np_dtypes(bool val) except+
void set_timestamp_type(data_type type) except+
void set_decimal_cols_as_float(vector[string] val) except+
void set_decimal128_columns(vector[string] val) except+
void enable_decimal128(bool val) except+

@staticmethod
orc_reader_options_builder builder(
Expand All @@ -58,9 +58,7 @@ cdef extern from "cudf/io/orc.hpp" \
orc_reader_options_builder& decimal_cols_as_float(
vector[string] val
) except+
orc_reader_options_builder& decimal128_columns(
vector[string] val
) except+
orc_reader_options_builder& decimal128(bool val) except+

orc_reader_options build() except+

Expand Down
1 change: 1 addition & 0 deletions python/cudf/cudf/_lib/orc.pyx
Original file line number Diff line number Diff line change
Expand Up @@ -248,6 +248,7 @@ cdef orc_reader_options make_orc_reader_options(
.timestamp_type(data_type(timestamp_type))
.use_index(use_index)
.decimal_cols_as_float(c_decimal_cols_as_float)
.decimal128(False)
.build()
)

Expand Down