Apache Arrow is a cross-language development platform for in-memory data. It specifies a standardized language-independent columnar memory format for flat and hierarchical data, organized for efficient analytic operations on modern hardware. It also provides computational libraries and zero-copy streaming messaging and interprocess communication.
The arrow package exposes an interface to the Arrow C++ library to access many of its features in R. This includes support for working with Parquet (read_parquet(), write_parquet()) and Feather (read_feather(), write_feather()) files, as well as lower-level access to Arrow memory and messages.
Install the latest release of arrow from CRAN with
install.packages("arrow")
Conda users on Linux and macOS can install arrow from conda-forge with
conda install -c conda-forge r-arrow
On macOS and Windows, installing a binary package from CRAN will handle Arrow’s C++ dependencies for you. On Linux, unless you use conda you’ll need to first install the C++ library. See the Arrow project installation page to find pre-compiled binary packages for some common Linux distributions, including Debian, Ubuntu, and CentOS. You’ll need to install libparquet-dev on Debian and Ubuntu, or parquet-devel on CentOS. This will also automatically install the Arrow C++ library as a dependency. Other Linux distributions must install the C++ library from source.
If you install the arrow package from source and the C++ library is not found, the R package functions will notify you that Arrow is not available. Call
arrow::install_arrow()
for version- and platform-specific guidance on installing the Arrow C++ library.
When installing from source, if the R and C++ library versions do not match, installation may fail. If you’ve previously installed the libraries and want to upgrade the R package, you’ll need to update the Arrow C++ library first.
library(arrow) set.seed(24) tab <- Table$create( x = 1:10, y = rnorm(10), z = as.factor(rep(c("b", "c"), 5)) ) tab #> Table #> 10 rows x 3 columns #> $x <int32> #> $y <double> #> $z <dictionary<values=string, indices=int8>> tab$x #> ChunkedArray #> <int32> #> [ #> 1, #> 2, #> 3, #> 4, #> 5, #> 6, #> 7, #> 8, #> 9, #> 10 #> ] as.data.frame(tab) #> x y z #> 1 1 -0.545880758 b #> 2 2 0.536585304 c #> 3 3 0.419623149 b #> 4 4 -0.583627199 c #> 5 5 0.847460017 b #> 6 6 0.266021979 c #> 7 7 0.444585270 b #> 8 8 -0.466495124 c #> 9 9 -0.848370044 b #> 10 10 0.002311942 c
Binary R packages for macOS and Windows are built daily and hosted at https://dl.bintray.com/ursalabs/arrow-r/. To install from there:
install.packages("arrow", repos="https://dl.bintray.com/ursalabs/arrow-r")
These daily package builds are not official Apache releases and are not recommended for production use. They may be useful for testing bug fixes and new features under active development.
Linux users will need to build the Arrow C++ library from source. See “Development” below. Once you have the C++ library, you can install the R package from GitHub using the remotes package. From within an R session,
# install.packages("remotes") # Or install "devtools", which includes remotes remotes::install_github("apache/arrow/r")
or if you prefer to stay at the command line,
R -e 'remotes::install_github("apache/arrow/r")'
You can specify a particular commit, branch, or release to install by including a ref argument to install_github(). This is particularly useful to match the R package version to the C++ library version you’ve installed.
Windows and macOS users who wish to contribute to the R package and don’t need to alter the Arrow C++ library may be able to obtain a recent version of the library without building from source. On macOS, you may install the C++ library using Homebrew:
# For the released version: brew install apache-arrow # Or for a development version, you can try: brew install apache-arrow --HEAD
On Windows, you can download a .zip file with the arrow dependencies from the rwinlib project, and then set the RWINLIB_LOCAL environment variable to point to that zip file before installing the arrow R package. That project contains released versions of the C++ library; for a development version, Windows users may be able to find a binary by going to the Apache Arrow project’s Appveyor, selecting an R job from a recent build, and downloading the build\arrow-*.zip file from the “Artifacts” tab.
If you need to alter both the Arrow C++ library and the R package code, or if you can’t get a binary version of the latest C++ library elsewhere, you’ll need to build it from source too.
First, install the C++ library. See the C++ developer guide for details.
Note that after any change to the C++ library, you must reinstall it and run make clean or git clean -fdx . to remove any cached object code in the r/src/ directory before reinstalling the R package. This is only necessary if you make changes to the C++ library source; you do not need to manually purge object files if you are only editing R or Rcpp code inside r/.
Once you’ve built the C++ library, you can install the R package and its dependencies, along with additional dev dependencies, from the git checkout:
cd ../../r R -e 'install.packages(c("devtools", "roxygen2", "pkgdown")); devtools::install_dev_deps()' R CMD INSTALL .
If you need to set any compilation flags while building the Rcpp extensions, you can use the ARROW_R_CXXFLAGS environment variable. For example, if you are using perf to profile the R extensions, you may need to set
export ARROW_R_CXXFLAGS=-fno-omit-frame-pointer
If the package fails to install/load with an error like this:
** testing if installed package can be loaded from temporary location Error: package or namespace load failed for 'arrow' in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '/Users/you/R/00LOCK-r/00new/arrow/libs/arrow.so': dlopen(/Users/you/R/00LOCK-r/00new/arrow/libs/arrow.so, 6): Library not loaded: @rpath/libarrow.14.dylib
try setting the environment variable R_LD_LIBRARY_PATH to wherever Arrow C++ was put in make install, e.g. export R_LD_LIBRARY_PATH=/usr/local/lib, and retry installing the R package.
For any other build/configuration challenges, see the C++ developer guide.
The arrow package uses some customized tools on top of Rcpp to prepare its C++ code in src/. If you change C++ code in the R package, you will need to set the ARROW_R_DEV environment variable to TRUE (optionally, add it to your~/.Renviron file to persist across sessions) so that the data-raw/codegen.R file is used for code generation.
The codegen.R script has these dependencies:
remotes::install_github("romainfrancois/decor") install.packages(c("dplyr", "purrr", "glue"))
We use Google C++ style in our C++ code. Check for style errors with
./lint.sh
Fix any style issues before committing with
./lint.sh --fix
The lint script requires Python 3 and clang-format-7. If the command isn’t found, you can explicitly provide the path to it like CLANG_FORMAT=$(which clang-format-7) ./lint.sh. On macOS, you can get this by installing LLVM via Homebrew and running the script as CLANG_FORMAT=$(brew --prefix llvm@7)/bin/clang-format ./lint.sh
Within an R session, these can help with package development:
devtools::load_all() # Load the dev package devtools::test(filter="^regexp$") # Run the test suite, optionally filtering file names devtools::document() # Update roxygen documentation rmarkdown::render("README.Rmd") # To rebuild README.md pkgdown::build_site() # To preview the documentation website devtools::check() # All package checks; see also below
Any of those can be run from the command line by wrapping them in R -e '$COMMAND'. There’s also a Makefile to help with some common tasks from the command line (make test, make doc, make clean, etc.)
R CMD build --keep-empty-dirs . R CMD check arrow_*.tar.gz --as-cran --no-manual