check_data: Check Data

In chk: Check User-Supplied Function Arguments

Check Data

Description

Checks column names, values, number of rows and key for a data.frame.

Usage

check_data(
  x,
  values = NULL,
  exclusive = FALSE,
  order = FALSE,
  nrow = numeric(0),
  key = character(0),
  x_name = NULL
)

Arguments

x	The object to check.
values	A uniquely named list of atomic vectors of the column values.
exclusive	A flag specifying whether x must only include columns named in values.
order	A flag specifying whether the order of columns in x must match names in values.
nrow	A flag or a whole numeric vector of the value, value range or possible values.
key	A character vector of the columns that represent a unique key.
x_name	A string of the name of object x or NULL.

Details

The values of each column are checked using check_values(). See its details for the rules including those for factor levels and missing values.

values is reduced with unique(), so a column whose class lacks a correct unique() method is checked against the wrong values. In particular unique() on a lubridate Period vector discards everything but the seconds, which collapses values and silently changes the check applied (see https://github.com/tidyverse/lubridate/issues/1154). Either define unique.Period() in your own code or convert the column and values to a class with a correct unique() method, such as the number of seconds with as.numeric(lubridate::as.duration(x)).

Value

An informative error if the test fails or an invisible copy of x.

See Also

Other check: check_dim(), check_dirs(), check_files(), check_key(), check_length(), check_names(), check_values()

Examples

check_data(data.frame())
check_data(data.frame(x = 2), list(x = 1))
try(check_data(data.frame(x = 2), list(y = 1L)))
try(check_data(data.frame(x = 2), list(y = 1)))
try(check_data(data.frame(x = 2), nrow = 2))

chk documentation built on Sept. 3, 2026, 9:07 a.m.