Read-Write Support for 'NumPy' Files via 'Rcpp'

The 'cnpy' library written by Carl Rogers provides read and write facilities for files created with (or for) the 'NumPy' extension for 'Python'. Vectors and matrices of numeric types can be read or written to and from files as well as compressed files. Support for integer files is available if the package has been built with -std=c++11 which is the default starting with release 0.2.3 following the release of R 3.1.0, and available on all platforms following the release of R 3.3.0 with the updated 'Rtools'.


This package uses the cnpy library written by Carl Rogers to provide read and write facilities for files created with (or for) the NumPy extension for Python. Vectors and matrices of numeric types can be read or written to and from files as well as compressed files. Support for integer files is available if the package has been built with -std=c++11 which is the default starting with release 0.2.3 following the release of R 3.1.0, and available on all platforms following the release of R 3.3.0 with the updated 'Rtools'.

The following Python code

>>> fm = np.arange(12).reshape(3,4) * 1.1
>>> fm
array([[  0. ,   1.1,   2.2,   3.3],
       [  4.4,   5.5,   6.6,   7.7],
       [  8.8,   9.9,  11. ,  12.1]])
>>> np.save("fmat.npy", fm)
>>> 
>>> im = np.arange(12).reshape(3,4)
>>> im
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> np.save("imat.npy", im)
>>> 

saves two matrices in floating-point and integer representation.

With this R code we can read and assign the files:

R> library(RcppCNPy)
R> fmat <- npyLoad("fmat.npy")
R> fmat
     [,1] [,2] [,3] [,4]
[1,]  0.0  1.1  2.2  3.3
[2,]  4.4  5.5  6.6  7.7
[3,]  8.8  9.9 11.0 12.1
R> 
R> imat <- npyLoad("imat.npy", "integer")
R> imat
     [,1] [,2] [,3] [,4]
[1,]    0    1    2    3
[2,]    4    5    6    7
[3,]    8    9   10   11
R> 

Going the opposite way by saving in R and reading in Python works equally well. An extension not present in CNPy allows reading and writing of gzip-compressed files.

The package has been tested and used on several architecture, and copes correctly with little-vs-big endian switches.

More details are available in the package vignette.

The package is on CRAN and can be installed per:

R> install.packages(RCppCNPy)

On CRAN, stable and mostly feature-complete.

Contributions are welcome, please use the GitHub issue tracker for bug reports, feature requests or general discussions before sending pull requests.

Dirk Eddelbuettel and Wush Wu

GPL (>= 2)

News

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("RcppCNPy")

0.2.7 by Dirk Eddelbuettel, 2 months ago


http://dirk.eddelbuettel.com/code/rcpp.cnpy.html


Report a bug at https://github.com/eddelbuettel/rcppcnpy/issues


Browse source code at https://github.com/cran/RcppCNPy


Authors: Dirk Eddelbuettel and Wush Wu


Documentation:   PDF Manual  


GPL (>= 2) license


Imports methods, Rcpp

Suggests knitr, rmarkdown, pinp

Linking to Rcpp


See at CRAN