use_data()
makes it easy to save package data in the correct format. I
recommend you save scripts that generate package data in data-raw
: use
use_data_raw()
to set it up. You also need to document exported datasets.
use_data(
...,
internal = FALSE,
overwrite = FALSE,
compress = "bzip2",
version = 2,
ascii = FALSE
)use_data_raw(name = "DATASET", open = rlang::is_interactive())
Unquoted names of existing objects to save.
If FALSE
, saves each object in its own .rda
file in the data/
directory. These data files bypass the usual
export mechanism and are available whenever the package is loaded
(or via data()
if LazyData
is not true).
If TRUE
, stores all objects in a single R/sysdata.rda
file.
Objects in this file follow the usual export rules. Note that this means
they will be exported if you are using the common exportPattern()
rule which exports all objects except for those that start with .
.
By default, use_data()
will not overwrite existing
files. If you really want to do so, set this to TRUE
.
Choose the type of compression used by save()
.
Should be one of "gzip", "bzip2", or "xz".
The serialization format version to use. The default, 2, was the default format from R 1.4.0 to 3.5.3. Version 3 became the default from R 3.6.0 and can only be read by R versions 3.5.0 and higher.
if TRUE
, an ASCII representation of the data is
written. The default value of ascii
is FALSE
which
leads to a binary file being written. If NA
and
version >= 2
, a different ASCII representation is used which
writes double/complex numbers as binary fractions.
Name of the dataset to be prepared for inclusion in the package.
Open the newly created file for editing? Happens in RStudio, if
applicable, or via utils::file.edit()
otherwise.
The data chapter of R Packages.
if (FALSE) {
x <- 1:10
y <- 1:100
use_data(x, y) # For external use
use_data(x, y, internal = TRUE) # For internal use
}
if (FALSE) {
use_data_raw("daisy")
}
Run the code above in your browser using DataLab