From 688cfcd24af3bb33ae0a55c466b0f41062f01353 Mon Sep 17 00:00:00 2001 From: Mashrfee Aryan Date: Fri, 9 Oct 2026 10:06:03 -0500 Subject: [PATCH 1/3] Add adaptive compression for Parquet V2 pages Add min_space_savings to control how much space compression must save. V2 pages below the threshold are written uncompressed. Keep the default behavior and V1 pages unchanged. --- cpp/src/parquet/column_writer.cc | 13 ++++++- cpp/src/parquet/column_writer_test.cc | 55 ++++++++++++++++++++++++++- cpp/src/parquet/properties.h | 24 +++++++++++- cpp/src/parquet/properties_test.cc | 27 +++++++++++++ 4 files changed, 114 insertions(+), 5 deletions(-) diff --git a/cpp/src/parquet/column_writer.cc b/cpp/src/parquet/column_writer.cc index 3296af62f0c4..9d8026c3cc08 100644 --- a/cpp/src/parquet/column_writer.cc +++ b/cpp/src/parquet/column_writer.cc @@ -1057,8 +1057,17 @@ void ColumnWriterImpl::BuildDataPageV2(int64_t definition_levels_rle_size, bool page_is_compressed = false; if (pager_->has_compressor() && values->size() > 0) { pager_->Compress(*values, compressor_temp_buffer_.get()); - if (compressor_temp_buffer_->size() < values->size()) { - page_is_compressed = true; + + if (const auto minimum = properties_->min_space_savings()) { + + //Checks if compressed size meets the minimum threshold + const double savings = + 1.0 - static_cast(compressor_temp_buffer_->size()) / values->size(); + page_is_compressed = savings >= *minimum; + } else { + // Keeps original behavior + // Allows users keep min_space_savings unset + page_is_compressed = compressor_temp_buffer_->size() < values->size(); } } std::shared_ptr compressed_values = diff --git a/cpp/src/parquet/column_writer_test.cc b/cpp/src/parquet/column_writer_test.cc index 69ec07c1b78f..a65facf4e8a4 100644 --- a/cpp/src/parquet/column_writer_test.cc +++ b/cpp/src/parquet/column_writer_test.cc @@ -15,7 +15,9 @@ // specific language governing permissions and limitations // under the License. +#include #include +#include #include #include @@ -112,7 +114,8 @@ class TestPrimitiveWriter : public PrimitiveTypedTest { const ParquetVersion::type version = ParquetVersion::PARQUET_1_0, const ParquetDataPageVersion data_page_version = ParquetDataPageVersion::V1, bool enable_checksum = false, int64_t page_size = kDefaultDataPageSize, - int64_t max_rows_per_page = kDefaultMaxRowsPerPage) { + int64_t max_rows_per_page = kDefaultMaxRowsPerPage, + std::optional min_space_savings = {}) { sink_ = CreateOutputStream(); WriterProperties::Builder wp_builder; wp_builder.version(version)->data_page_version(data_page_version); @@ -130,6 +133,7 @@ class TestPrimitiveWriter : public PrimitiveTypedTest { wp_builder.max_statistics_size(column_properties.max_statistics_size()); wp_builder.data_pagesize(page_size); wp_builder.max_rows_per_page(max_rows_per_page); + wp_builder.min_space_savings(min_space_savings); writer_properties_ = wp_builder.build(); metadata_ = ColumnChunkMetaDataBuilder::Make(writer_properties_, this->descr_); @@ -2084,6 +2088,55 @@ TEST_F(TestValuesWriterInt32Type, AvoidCompressedInDataPageV2) { verify_only_one_uncompressed_page(/*total_num_values=*/1); } } + +// Tests if compression works when: +// - min_space_savings is unset +// - min_space_savings = 0.1 +// - min_space_savings = 1 +TEST_F(TestValuesWriterInt32Type, MinSpaceSavingsDataPageV2) { + + // Used zeros because it compresses well + this->SetUpSchema(Repetition::OPTIONAL); + this->descr_ = this->schema_.Column(0); + this->GenerateData(SMALL_SIZE); + std::fill(this->values_.begin(), this->values_.end(), 0); + + ColumnProperties column_properties; + column_properties.set_compression(Compression::ZSTD); + + const std::vector, bool>> cases = { + {std::nullopt, true}, + {0.1, true}, + {1.0, false}, + }; + + for (const auto& [minimum, expected_compressed] : cases) { + auto writer = this->BuildWriter( + SMALL_SIZE, column_properties, ParquetVersion::PARQUET_2_LATEST, + ParquetDataPageVersion::V2, false, kDefaultDataPageSize, + kDefaultMaxRowsPerPage, minimum); + + writer->WriteBatch(SMALL_SIZE, this->def_levels_.data(), nullptr, + this->values_ptr_); + writer->Close(); + + ASSERT_OK_AND_ASSIGN(auto buffer, this->sink_->Finish()); + + auto page_reader = PageReader::Open( + std::make_shared<::arrow::io::BufferReader>(buffer), SMALL_SIZE, + Compression::ZSTD, default_reader_properties(), *this->descr_); + + auto page = page_reader->NextPage(); + + ASSERT_NE(page, nullptr); + ASSERT_EQ(PageType::DATA_PAGE_V2, page->type()); + + auto data_page = std::static_pointer_cast(page); + + ASSERT_EQ(expected_compressed, data_page->is_compressed()); + } +} + #endif // Test writing and reading geometry columns diff --git a/cpp/src/parquet/properties.h b/cpp/src/parquet/properties.h index 89f2be42bffd..3bb721c7b560 100644 --- a/cpp/src/parquet/properties.h +++ b/cpp/src/parquet/properties.h @@ -403,6 +403,7 @@ class PARQUET_EXPORT WriterProperties { max_rows_per_page_(properties.max_rows_per_page()), version_(properties.version()), data_page_version_(properties.data_page_version()), + min_space_savings_(properties.min_space_savings()), created_by_(properties.created_by()), store_decimal_as_integer_(properties.store_decimal_as_integer()), page_checksum_enabled_(properties.page_checksum_enabled()), @@ -529,6 +530,13 @@ class PARQUET_EXPORT WriterProperties { return this; } + /// Used to set minimum threshold for V2 page data values compression + /// If min_space_savings unset, keep compression if compressed value is smaller. + Builder* min_space_savings(std::optional min_space_savings) { + min_space_savings_ = min_space_savings; + return this; + } + /// Specify the Parquet file version. /// Default PARQUET_2_6. Builder* version(ParquetVersion::type version) { @@ -883,6 +891,11 @@ class PARQUET_EXPORT WriterProperties { /// \brief Build the WriterProperties with the builder parameters. /// \return The WriterProperties defined by the builder. std::shared_ptr build() { + // checks for invalid input of min_space_savings + if (min_space_savings_ && + !(*min_space_savings_ >= 0.0 && *min_space_savings_ <= 1.0)) { + throw ParquetException("min_space_savings must be in range [0, 1]"); + } std::unordered_map column_properties; auto get = [&](const std::string& key) -> ColumnProperties& { auto it = column_properties.find(key); @@ -918,7 +931,7 @@ class PARQUET_EXPORT WriterProperties { pagesize_, max_rows_per_page_, version_, created_by_, page_checksum_enabled_, size_statistics_level_, std::move(file_encryption_properties_), default_column_properties_, column_properties, data_page_version_, - store_decimal_as_integer_, std::move(sorting_columns_), + min_space_savings_, store_decimal_as_integer_, std::move(sorting_columns_), content_defined_chunking_enabled_, content_defined_chunking_options_)); } @@ -933,6 +946,7 @@ class PARQUET_EXPORT WriterProperties { int64_t max_rows_per_page_; ParquetVersion::type version_; ParquetDataPageVersion data_page_version_; + std::optional min_space_savings_ = {}; std::string created_by_; bool store_decimal_as_integer_; bool page_checksum_enabled_; @@ -973,6 +987,9 @@ class PARQUET_EXPORT WriterProperties { return parquet_data_page_version_; } + + std::optional min_space_savings() const { return min_space_savings_; } + inline ParquetVersion::type version() const { return parquet_version_; } inline std::string created_by() const { return parquet_created_by_; } @@ -1100,7 +1117,8 @@ class PARQUET_EXPORT WriterProperties { std::shared_ptr file_encryption_properties, const ColumnProperties& default_column_properties, const std::unordered_map& column_properties, - ParquetDataPageVersion data_page_version, bool store_short_decimal_as_integer, + ParquetDataPageVersion data_page_version, std::optional min_space_savings, + bool store_short_decimal_as_integer, std::vector sorting_columns, bool content_defined_chunking_enabled, CdcOptions content_defined_chunking_options) : pool_(pool), @@ -1110,6 +1128,7 @@ class PARQUET_EXPORT WriterProperties { pagesize_(pagesize), max_rows_per_page_(max_rows_per_page), parquet_data_page_version_(data_page_version), + min_space_savings_(min_space_savings), parquet_version_(version), parquet_created_by_(created_by), store_decimal_as_integer_(store_short_decimal_as_integer), @@ -1129,6 +1148,7 @@ class PARQUET_EXPORT WriterProperties { int64_t pagesize_; int64_t max_rows_per_page_; ParquetDataPageVersion parquet_data_page_version_; + std::optional min_space_savings_; ParquetVersion::type parquet_version_; std::string parquet_created_by_; bool store_decimal_as_integer_; diff --git a/cpp/src/parquet/properties_test.cc b/cpp/src/parquet/properties_test.cc index 2f22844f8bbf..f56d38eba619 100644 --- a/cpp/src/parquet/properties_test.cc +++ b/cpp/src/parquet/properties_test.cc @@ -80,6 +80,33 @@ TEST(TestWriterProperties, DefaultCompression) { ::arrow::util::kUseDefaultCompressionLevel); } +// checks if WriterProperties correctly store, copy, and clear min_space_savings +TEST(TestWriterProperties, MinSpaceSavings) { + WriterProperties::Builder builder; + ASSERT_FALSE(builder.build()->min_space_savings().has_value()); + + for (double savings : {0.0, 0.1, 1.0}) { + auto properties = builder.min_space_savings(savings)->build(); + ASSERT_EQ(std::optional(savings), properties->min_space_savings()); + + auto copied = WriterProperties::Builder(*properties).build(); + ASSERT_EQ(properties->min_space_savings(), copied->min_space_savings()); + } + + auto properties = builder.min_space_savings(std::nullopt)->build(); + ASSERT_FALSE(properties->min_space_savings().has_value()); +} + +// Tests for invalid settings +TEST(TestWriterProperties, InvalidMinSpaceSavings) { + for (double savings : {-0.1, 1.1, std::numeric_limits::infinity(), + -std::numeric_limits::infinity(), + std::numeric_limits::quiet_NaN()}) { + EXPECT_THROW(WriterProperties::Builder().min_space_savings(savings)->build(), + ParquetException); + } +} + TEST(TestWriterProperties, AdvancedHandling) { WriterProperties::Builder builder; builder.compression("gzip", Compression::GZIP); From 01512cc6cec1e2bc1db21bd7eb40b1f201d72af8 Mon Sep 17 00:00:00 2001 From: Mashrfee Aryan Date: Fri, 9 Oct 2026 11:37:50 -0500 Subject: [PATCH 2/3] Let Python users set minimum compression savings Add an option to choose how much space compression must save when writing Parquet files from Python. Add tests as well. --- python/pyarrow/_dataset_parquet.pyx | 2 + python/pyarrow/_parquet.pxd | 3 +- python/pyarrow/_parquet.pyx | 16 +++++-- python/pyarrow/includes/libparquet.pxd | 1 + python/pyarrow/parquet/core.py | 4 ++ .../tests/parquet/test_parquet_writer.py | 44 +++++++++++++++++++ python/pyarrow/tests/test_dataset.py | 21 +++++++++ 7 files changed, 87 insertions(+), 4 deletions(-) diff --git a/python/pyarrow/_dataset_parquet.pyx b/python/pyarrow/_dataset_parquet.pyx index 43eca8957be8..ffff69f4bcfe 100644 --- a/python/pyarrow/_dataset_parquet.pyx +++ b/python/pyarrow/_dataset_parquet.pyx @@ -655,6 +655,7 @@ cdef class ParquetFileWriteOptions(FileWriteOptions): ), column_encoding=self._properties["column_encoding"], data_page_version=self._properties["data_page_version"], + min_space_savings=self._properties["min_space_savings"], encryption_properties=self._properties["encryption_properties"], write_batch_size=self._properties["write_batch_size"], dictionary_pagesize_limit=self._properties["dictionary_pagesize_limit"], @@ -707,6 +708,7 @@ cdef class ParquetFileWriteOptions(FileWriteOptions): use_byte_stream_split=False, column_encoding=None, data_page_version="1.0", + min_space_savings=None, use_deprecated_int96_timestamps=False, coerce_timestamps=None, allow_truncated_timestamps=False, diff --git a/python/pyarrow/_parquet.pxd b/python/pyarrow/_parquet.pxd index 36fc2ccf2f33..28a5c676ec98 100644 --- a/python/pyarrow/_parquet.pxd +++ b/python/pyarrow/_parquet.pxd @@ -57,7 +57,8 @@ cdef shared_ptr[WriterProperties] _create_writer_properties( sorting_columns=*, store_decimal_as_integer=*, use_content_defined_chunking=*, - bloom_filter_options=* + bloom_filter_options=*, + min_space_savings=* ) except * diff --git a/python/pyarrow/_parquet.pyx b/python/pyarrow/_parquet.pyx index 2621c3bd6d61..640a69a2453d 100644 --- a/python/pyarrow/_parquet.pyx +++ b/python/pyarrow/_parquet.pyx @@ -2076,13 +2076,21 @@ cdef shared_ptr[WriterProperties] _create_writer_properties( sorting_columns=None, store_decimal_as_integer=False, use_content_defined_chunking=False, - bloom_filter_options=None) except *: + bloom_filter_options=None, + min_space_savings=None) except *: """General writer properties""" cdef: shared_ptr[WriterProperties] properties WriterProperties.Builder props CdcOptions cdc_options + double savings + + if min_space_savings is not None: + savings = min_space_savings + if not (0.0 <= savings <= 1.0): + raise ValueError("min_space_savings must be in range [0, 1]") + props.min_space_savings(optional[double](savings)) # data_page_version @@ -2405,7 +2413,8 @@ cdef class ParquetWriter(_Weakrefable): store_decimal_as_integer=False, use_content_defined_chunking=False, write_time_adjusted_to_utc=False, - bloom_filter_options=None): + bloom_filter_options=None, + min_space_savings=None): cdef: shared_ptr[WriterProperties] properties shared_ptr[ArrowWriterProperties] arrow_properties @@ -2442,7 +2451,8 @@ cdef class ParquetWriter(_Weakrefable): sorting_columns=sorting_columns, store_decimal_as_integer=store_decimal_as_integer, use_content_defined_chunking=use_content_defined_chunking, - bloom_filter_options=bloom_filter_options + bloom_filter_options=bloom_filter_options, + min_space_savings=min_space_savings, ) arrow_properties = _create_arrow_writer_properties( use_deprecated_int96_timestamps=use_deprecated_int96_timestamps, diff --git a/python/pyarrow/includes/libparquet.pxd b/python/pyarrow/includes/libparquet.pxd index 3a5f91716c17..c95d81a26761 100644 --- a/python/pyarrow/includes/libparquet.pxd +++ b/python/pyarrow/includes/libparquet.pxd @@ -485,6 +485,7 @@ cdef extern from "parquet/api/writer.h" namespace "parquet" nogil: cdef cppclass WriterProperties: cppclass Builder: Builder* data_page_version(ParquetDataPageVersion version) + Builder* min_space_savings(optional[double] min_space_savings) Builder* version(ParquetVersion version) Builder* compression(ParquetCompression codec) Builder* compression(const c_string& path, diff --git a/python/pyarrow/parquet/core.py b/python/pyarrow/parquet/core.py index 4de88fe22997..bccf47b917ed 100644 --- a/python/pyarrow/parquet/core.py +++ b/python/pyarrow/parquet/core.py @@ -1090,6 +1090,7 @@ def __init__(self, where, schema, filesystem=None, max_rows_per_page=None, bloom_filter_options=None, use_content_defined_chunking=False, + min_space_savings=None, **options): if use_deprecated_int96_timestamps is None: # Use int96 timestamps for Spark @@ -1147,6 +1148,7 @@ def __init__(self, where, schema, filesystem=None, max_rows_per_page=max_rows_per_page, bloom_filter_options=bloom_filter_options, use_content_defined_chunking=use_content_defined_chunking, + min_space_savings=min_space_savings, **options) self.is_open = True @@ -2039,6 +2041,7 @@ def write_table(table, where, row_group_size=None, version='2.6', max_rows_per_page=None, bloom_filter_options=None, use_content_defined_chunking=False, + min_space_savings=None, **kwargs): # Implementor's note: when adding keywords here / updating defaults, also # update it in write_to_dataset and _dataset_parquet.pyx ParquetFileWriteOptions @@ -2074,6 +2077,7 @@ def write_table(table, where, row_group_size=None, version='2.6', max_rows_per_page=max_rows_per_page, bloom_filter_options=bloom_filter_options, use_content_defined_chunking=use_content_defined_chunking, + min_space_savings=min_space_savings, **kwargs) as writer: writer.write_table(table, row_group_size=row_group_size) except Exception: diff --git a/python/pyarrow/tests/parquet/test_parquet_writer.py b/python/pyarrow/tests/parquet/test_parquet_writer.py index a49441f09f45..89b8eca78f6d 100644 --- a/python/pyarrow/tests/parquet/test_parquet_writer.py +++ b/python/pyarrow/tests/parquet/test_parquet_writer.py @@ -41,6 +41,50 @@ pytestmark = pytest.mark.parquet +@pytest.mark.parametrize("minimum", [None, 0.0, 0.1, 1.0]) +@pytest.mark.parametrize("page_version", ["1.0", "2.0"]) +@pytest.mark.parametrize("compression", ["none", "snappy"]) +@pytest.mark.parametrize("use_writer", [False, True]) +def test_min_space_savings( + minimum, page_version, compression, use_writer): + table = pa.table({ + "value": pa.array([0.0] * 1000, type=pa.float32()) + }) + sink = pa.BufferOutputStream() + options = dict( + compression=compression, + use_dictionary=False, + data_page_version=page_version, + min_space_savings=minimum, + ) + + if use_writer: + with pq.ParquetWriter(sink, table.schema, **options) as writer: + writer.write_table(table) + else: + pq.write_table(table, sink, **options) + + assert pq.read_table(sink.getvalue()).equals(table) + +# Tests invalid settings +@pytest.mark.parametrize( + "minimum", + [-0.1, 1.1, float("inf"), -float("inf"), float("nan")] +) +@pytest.mark.parametrize("use_writer", [False, True]) +def test_invalid_min_space_savings(minimum, use_writer): + table = pa.table({"value": [1.0]}) + sink = pa.BufferOutputStream() + + with pytest.raises(ValueError, match="min_space_savings"): + if use_writer: + pq.ParquetWriter( + sink, table.schema, min_space_savings=minimum) + else: + pq.write_table( + table, sink, min_space_savings=minimum) + + @pytest.mark.pandas def test_parquet_incremental_file_build(tempdir): df = _test_dataframe(100) diff --git a/python/pyarrow/tests/test_dataset.py b/python/pyarrow/tests/test_dataset.py index 0a94c0bd9875..7daf2dfb2558 100644 --- a/python/pyarrow/tests/test_dataset.py +++ b/python/pyarrow/tests/test_dataset.py @@ -5949,6 +5949,27 @@ def test_write_dataset_content_defined_chunking(tempdir, cdc): assert ds.dataset(tempdir, format=fmt).to_table().equals(expected_table) +@pytest.mark.parquet +@pytest.mark.parametrize("minimum", [None, 0.0, 0.1, 1.0]) +def test_parquet_min_space_savings(tempdir, minimum): + import pyarrow.parquet as pq + + table = pa.table({ + "value": pa.array([0.0] * 1000, type=pa.float32()) + }) + pq.write_to_dataset( + table, + tempdir, + compression="snappy", + use_dictionary=False, + data_page_version="2.0", + min_space_savings=minimum, + ) + + result = ds.dataset(tempdir, format="parquet").to_table() + assert result.equals(table) + + def test_make_write_options_error(): # GH-39440: calling make_write_options as a static class method msg_1 = ("make_write_options() should be called on an " From 89bf1b25ce6d4c6f667a86de309a78625dda1e35 Mon Sep 17 00:00:00 2001 From: Mashrfee Aryan Date: Sat, 10 Oct 2026 04:54:01 -0500 Subject: [PATCH 3/3] Fix ignored compression savings in dataset writing Dataset writing ignored min_space_savings. Move the argument to the end of the call so it reaches the writer. Add a test to check that the setting changes the file size. --- python/pyarrow/_dataset_parquet.pyx | 2 +- python/pyarrow/tests/test_dataset.py | 20 ++++++++++++++++++++ 2 files changed, 21 insertions(+), 1 deletion(-) diff --git a/python/pyarrow/_dataset_parquet.pyx b/python/pyarrow/_dataset_parquet.pyx index ffff69f4bcfe..4b9d2e2cc8ee 100644 --- a/python/pyarrow/_dataset_parquet.pyx +++ b/python/pyarrow/_dataset_parquet.pyx @@ -655,7 +655,6 @@ cdef class ParquetFileWriteOptions(FileWriteOptions): ), column_encoding=self._properties["column_encoding"], data_page_version=self._properties["data_page_version"], - min_space_savings=self._properties["min_space_savings"], encryption_properties=self._properties["encryption_properties"], write_batch_size=self._properties["write_batch_size"], dictionary_pagesize_limit=self._properties["dictionary_pagesize_limit"], @@ -665,6 +664,7 @@ cdef class ParquetFileWriteOptions(FileWriteOptions): store_decimal_as_integer=self._properties["store_decimal_as_integer"], use_content_defined_chunking=self._properties["use_content_defined_chunking"], bloom_filter_options=self._properties["bloom_filter_options"], + min_space_savings=self._properties["min_space_savings"], ) def _set_arrow_properties(self): diff --git a/python/pyarrow/tests/test_dataset.py b/python/pyarrow/tests/test_dataset.py index 7daf2dfb2558..589bc08f037c 100644 --- a/python/pyarrow/tests/test_dataset.py +++ b/python/pyarrow/tests/test_dataset.py @@ -5970,6 +5970,26 @@ def test_parquet_min_space_savings(tempdir, minimum): assert result.equals(table) +@pytest.mark.parquet +def test_dataset_min_space_savings_changes_size(tempdir): + table = pa.table({"value": [0.0] * 1000}) + sizes = [] + + for minimum in [None, 1.0]: + directory = tempdir / str(minimum) + pq.write_to_dataset( + table, + directory, + compression="snappy", + use_dictionary=False, + data_page_version="2.0", + min_space_savings=minimum, + ) + sizes.append(next(directory.iterdir()).stat().st_size) + + assert sizes[1] > sizes[0] + + def test_make_write_options_error(): # GH-39440: calling make_write_options as a static class method msg_1 = ("make_write_options() should be called on an "