Back to Basics with Tibbles and data.tables

Sometimes we might wish for the old days be back and want to work with familiar objects. This function helps to convert tibbles and data.tables to plain data.frames as smoothly as possible.

Usage

toBaseR(x, ...)

# S3 method for class 'tbl_df'
toBaseR(x, ...)

# S3 method for class 'data.table'
toBaseR(x, ...)

# S3 method for class 'haven_labelled'
toBaseR(x, ...)

# Default S3 method
toBaseR(x, ...)

Arguments

x: the object to be converted.
...: arguments passed on.

Value

converted object.

Details

For a data.table the key ("sorted"), secondary indices ("index") and the ".internal.selfref" pointer are dropped, and a deep copy is returned, so that later modifications by reference on the original cannot leak into the result. Columns of class haven_labelled are converted to factors in both cases.

Examples

# a tibble is rolled back to a plain data.frame
if (requireNamespace("tibble", quietly = TRUE)) {
  tbl <- tibble::as_tibble(head(iris))
  class(toBaseR(tbl))
}
#> [1] "data.frame"

# same for a keyed data.table
dt <- data.table::as.data.table(head(iris))
data.table::setkey(dt, Species)
df <- toBaseR(dt)
class(df)
#> [1] "data.frame"
attr(df, "sorted")
#> NULL

# an object without a method is returned unchanged, with a warning
x <- suppressWarnings(toBaseR(1:3))
identical(x, 1:3)
#> [1] TRUE

# \donttest{
# labelled data from other statistical packages: needs 'haven' and
# an internet connection, hence the try()
if (requireNamespace("haven", quietly = TRUE)) {
  url <- "http://www.stata.com/videos13/data/webclass.dta"
  d.webclass <- try(toBaseR(haven::read_dta(url)))
}
# }