apache/arrow

Describe the bug The R bindings do not expose page_checksum_verification, so an R user has no way to turn on page-CRC verification when reading Parquet. Arrow's own deliberately-corrupt test files are read without error. C++ has the option and Python exposes it; only R does not. C++ cpp/src/parquet/properties.h:148-150 — page_checksum_verification() / set_page_checksum_verification(bool check_crc), defaulting to false at line 166 Python python/pyarrow/parquet/core.py:269,325,346 — a documented page_checksum_verification=False parameter R r/R/parquet.R:656-682 — ParquetReaderProperties declares and implements only thrift_string_size_limit and thrift_container_size_limit Reproduction Using the files parquet-testing ships expressly so readers can be shown to reject a bad CRC — datapage_v1-corrupt-checksum.parquet ("uncompressed INT32 columns in v1 data pages with a mismatching CRC") and rle-dict-uncompressed-corrupt-checksum.parquet: library(arrow) packageVersion("

Describe the enhancement requested Arrow C++ 26.0.0 adds two new arrow::csv::ParseOptions fields: ignore_extra_columns (#51118): rows with more columns than expected have the extra columns dropped instead of erroring. pad_short_rows (#50926): rows with fewer columns than expected are padded with nulls instead of erroring. Neither can be set from R. csv_parse_options() and CsvParseOptions$create() should accept both, and csv.cpp should pass them through: arrow/r/R/csv.R Lines 735 to 753 in 9479e68 quote_char = '"', double_quote = TRUE, escaping = FALSE, escape_char = "\\", newlines_in_values = FALSE, ignore_empty_lines = TRUE ) {

Describe the enhancement requested In the Python Parquet reader docstrings, many snippets are duplicated for describing various common options, such as thrift_string_size_limit. We should deduplicate these snippets and find a way to reuse them across all Parquet reader docstrings. This will ease maintenance and reduce the risk of errors when updating those docstrings. Component(s) Parquet, Python

Describe the enhancement requested In visit_type_inline.h we have a VisitType that dispatches all existing data types to a generic callable (presumably a lambda or template function). It would be useful to have a similar VisitIntegerType helper that assumes its input is an integer type. This would help implementing helpers for dictionary indices, take indices, etc. For example this snippet could be reimplemented more tersely using said helper: arrow/cpp/src/arrow/util/dict_util.cc Lines 61 to 79 in 154962c const auto& dict_array_type = internal::checked_cast<const DictionaryType&>(*span.type); switch (dict_array_type.index_type()->id()) { case Type::UINT8: return LogicalNullCount<UInt8Type>(span); case Type::INT8:

Describe the bug, including details regarding any error messages, version, and platform. The r-devdocs crossbow job consistently fails during the period between when the main Apache Arrow release completes (and C++ packages are published to apt) and when the R package is released to CRAN. Example failure from 23.0.1: https://github.com/ursacomputing/crossbow/actions/runs/22147872099/job/64030288380 The error is an ABI mismatch: Error: package or namespace load failed for 'arrow' in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '/home/runner/work/_temp/Library/arrow/libs/arrow.so': /home/runner/work/crossbow/crossbow/dist/lib/libarrow_compute.so.2300: undefined symbol: _ZN5arrow6Status14AddContextLineEPKciS2_ Root cause The r/vignettes/developers/setup.Rmd file contains a hidden test section (lines 342-351) that: Builds Arrow C++ from source (from the git checkout) Installs the R package linked against this custom-built libarrow Installs system libarrow-dev f