The checkr packages provides a series of functions that
can be used to check the values and classes of common R objects. Before
attempting to use the functions it is worth understanding the checkr
function naming schemes.
Primary Objects
As discussed in the Data Structures
chapter in Advanced R by Hadley Wickham, the five main object types are
defined by their dimensionality (1d, 2d and nd) and whether they are
homogenous or heterogenous. Currently the check_vector(),
check_list() and check_data() functions allow
the user to check the type and values of the three main type of (atomic)
vectors, lists and data frames. The check_matrix() and
check_array() equivalents have yet to be implemented. The
functions are named according to the type of the object they check.
It is worth noting that a while a vector, matrix or array will
definitely pass the check_homogenous() function, a list or
data frame may pass it if all its elements have identical classes.
Scalars
In addition to (atomic) vectors the checkr package also identifies
scalars which in R are (atomic) vectors with one element. They can be
tested using the check_scalar() function.
Often function arguments are scalars of a particular type. For
example the default mean() function has an argument
na.rm which is a logical indicating whether NA
values should be stripped. The checkr package provides the following
functions to test whether objects are non-missing scalars of particular
types:
-
check_lgl(): a non-missing logical scalar i.e.TRUEorFALSE -
check_int(): an non-missing integer i.e. 1L -
check_dbl(): an non-missing double scalar i.e. 1.1 -
check_prop(): a non-missing double scalar between zero and one, i.e. 0.35 -
check_chr(): a non-missing character scalar, i.e."text" -
check_date(): a non-missing Date scalar i.e.as.Date("2001-02-03") -
check_dttm(): a non-missing POSIXct scalar
It is also worth noting that almost all the scalar functions include
an argument coerce which if set to TRUE tests
whether the object satisfies the condition after coercing to the
required type. If the test passes then the function returns the coerced
object.
Other Objects
Functions to test whether an object is another type of object include
check_null(), check_environment() and
check_function().
In addition the check_classes() and
check_inherits() functions can be used to check the classes
or inheritance of an object.
Function Functions
Many other checking functions are named according to the main base
function that they use. Thus the check_nrow() function uses
the nrow() function to check the number of rows while the
check_nlevels() function uses the nlevels()
function. It is important to realize that these functions do not check
the type of an object but simply whether the return value of the base
function(s) is consistent with the user defined conditions.
Functions which fall in this category include
check_class(), check_colnames(),
check_length(), check_levels(),
check_names(), check_nchar(),
check_ncol(), check_nlevels(),
check_nrow() and check_sorted().
Data Functions
The check_join() function tests whether the columns in
one data frame form a many-to-one join with corresponding columns in a
second data frame. The check_key() function tests whether
specific columns are unique. They are useful for testing whether a set
of data frames can be archived in an relational database.
Miscellaneous Functions
The remaining functions which are not captured by the above naming
schemes include checkor(), check_named(),
check_pattern(), check_regex(),
check_sorted(), check_tzone(),
check_unique(), check_unnamed() and
check_unused().