README.md

In krlmlr/dfunbind: Column-wise loading of large data frames for package authors

dfunbind

When packaging data frames to a package, the entire data frame is loaded when first accessed. This is impractical if the data frame contains many columns and only few of them are needed. This package offers a workaround by providing functions to store and load data frames in one RDS file per column.

Installation

devtools::install_github("krlmlr/dfunbind")

Usage

To package a data set, use unbind(dataset, "inst/extdata/dataset"). Then, in the R/ subdirectory of your package, create a source file with the following contents:

#' @importFrom dfunbind isplice
#' @export
dataset <- isplice("dataset")

Use roxygen2 or devtools to create documentation and the NAMESPACE file. The dfunbind package needs to be imported explicitly to access S3 methods and operators defined there.

Example

The Iris dataset is packaged here as an example. Packaging happens by manually running data-raw/iris.R, it is made available to users of the package by R/iris.R. The data files are stored in inst/extdata/iris.

krlmlr/dfunbind documentation built on May 20, 2019, 6:16 p.m.