chk Families

Introduction

The vld_ functions are used within the chk_ functions. The chk_ functions (and their vld_ equivalents) can be divided into the following families.

In the code in this examples, we will use vld_* functions

If you want to learn more about the logic behind some of the functions explained here, we recommend reading the book Advanced R (Wickham, 2019).

For reasons of space, the x_name = NULL argument is not shown. For a more simplified list of the chk functions, you can see the Reference section.

chk_ Functions

Overview

Classification of the chk functions by family

Missing Input Checker

Check if the function input is missing or not

chk_missing function uses missing() to check if an argument has been left out when the function is called.

Function Code
chk_missing() missing()
chk_not_missing() !missing()

... Checker

Check if the function input comes from ... (dot-dot-dot) or not

The functions chk_used(...) and chk_unused(...) check if any arguments have been provided through ... (called dot-dot-dot or ellipsis), which is commonly used in R to allow a variable number of arguments.

Function Code
chk_used(...) length(list(...)) != 0L
chk_unused(...) length(list(...)) == 0L

External Data Source Checkers

Check if the function input is a valid external data source.

These chk functions check the existence of a file, the validity of its extension, and the existence of a directory.

Function Code
chk_file(x) vld_string(x) && file.exists(x) && !dir.exists(x)
chk_ext(x, ext) vld_string(x) && vld_subset(tools::file_ext(x), ext)
chk_dir(x) vld_string(x) && dir.exists(x)

NULL checker

Check if the function input is NULL or not

Function Code
chk_null(x) is.null(x)
chk_not_null(x) !is.null(x)

Scalar Checkers

Check if the function input is a scalar. In R, scalars are vectors of length 1.

Function Code
chk_scalar(x) length(x) == 1L

The following functions check if the functions inputs are vectors of length 1 of a particular data type. Each data type has a special syntax to create an individual value or “scalar”.

Function Code
chk_string(x) is.character(x) && length(x) == 1L && !anyNA(x)
chk_number(x) is.numeric(x) && length(x) == 1L && !anyNA(x)

For logical data types, you can check flags using chk_flag(), which considers TRUE or FALSE as possible values, or use chk_lgl() to verify if a scalar is of type logical, including NA as element.

Function Code
chk_flag(x) is.logical(x) && length(x) == 1L && !anyNA(x)
chk_lgl(x) is.logical(x) && length(x) == 1L

It is also possible to check if the user-provided argument is only TRUE or only FALSE:

Function Code
chk_true(x) is.logical(x) && length(x) == 1L && !anyNA(x) && x
chk_false(x) is.logical(x) && length(x) == 1L && !anyNA(x) && !x

Date or DateTime Checkers

Check if the function input is of class Date or DateTime

Date and datetime classes can be checked with chk_date and chk_datetime.

Function Code
chk_date(x) inherits(x, "Date") && length(x) == 1L && !anyNA(x)
chk_date_time(x) inherits(x, "POSIXct") && length(x) == 1L && !anyNA(x)

Time Zone Checker

Also you can check the time zone with chk_tz(). The available time zones can be retrieved using the function OlsonNames().

Function Code
chk_tz(x) is.character(x) && length(x) == 1L && !anyNA(x) && x %in% OlsonNames()

Data Structure Checker

Check if the function input has a specific data structure.

Vectors are a family of data types that come in two forms: atomic vectors and lists. When vectors consist of elements of the same data type, they can be considered atomic, matrices, or arrays. The elements in a list, however, can be of different types.

To check if a function argument is a vector you can use chk_vector().

Function Code
chk_vector(x) is.atomic(x) && !is.matrix(x) && !is.array(x)) || is.list(x)

Pay attention that chk_vector() and vld_vector() are different from is.vector(), that will return FALSE if the vector has any attributes except names.

vector <- c(1, 2, 3)
is.vector(vector) # TRUE
#> [1] TRUE
vld_vector(vector) # TRUE
#> [1] TRUE

attributes(vector) <-  list("a" = 10, "b" = 20, "c" = 30)
is.vector(vector) # FALSE
#> [1] FALSE
vld_vector(vector) # TRUE
#> [1] TRUE
Function Code
chk_atomic(x) is.atomic(x)

Notice that is.atomic is true for the types logical, integer, numeric, complex, character and raw. Also, it is TRUE for NULL.

vector <- c(1, 2, 3)
is.atomic(vector) # TRUE
#> [1] TRUE
vld_vector(vector) # TRUE
#> [1] TRUE

is.atomic(NULL) # TRUE
#> [1] FALSE
vld_vector(NULL) # TRUE
#> [1] FALSE

The dimension attribute converts vectors into matrices and arrays.

Function Code
chk_array(x) is.array(x)
chk_matrix(x) is.matrix(x)

When a vector is composed by heterogeneous data types, can be a list. Data frames are among the most important S3 vectors, constructed on top of lists.

Function Code
chk_list(x) is.list()
chk_data(x) inherits(x, "data.frame")

Be careful not to confuse the function chk_data with check_data. Please read the check_ functions section below and the function documentation.

Data Type Checkers

Check if the function input has a data type. You can use the function typeof() to confirm the data type.

Function Code
chk_environment(x) is.environment(x)
chk_logical(x) is.logical(x)
chk_character(x) is.character(x)

For numbers there are four functions. R differentiates between doubles (chk_double()) and integers (chk_integer()). You can also use the generic function chk_numeric(), which will detect both. The third type of number is complex (chk_complex()).

Function Code
chk_numeric(x) is.numeric(x)
chk_double(x) is.double(x)
chk_integer(x) is.integer(x)
chk_complex(x) is.complex(x)

Consider that to explicitly create an integer in R, you need to use the suffix L.

vld_numeric(33) # TRUE
#> [1] TRUE

vld_double(33) # TRUE
#> [1] TRUE
vld_integer(33) # FALSE
#> [1] FALSE

vld_integer(33L) # TRUE
#> [1] TRUE

Whole Number Checkers

These functions accept whole numbers, whether they are explicitly integers or double types without fractional parts.

Function Code
chk_whole_numeric is.integer(x) || (is.double(x) && vld_true(all.equal(x[!is.na(x)], trunc(x[!is.na(x)]))))
chk_whole_number vld_number(x) && (is.integer(x) || vld_true(all.equal(x, trunc(x))))
chk_count vld_whole_number(x) && x >= 0

If you want to consider both 3.0 and 3L as integers, it is safer to use the function chk_whole_numeric. Here, x is valid if it’s an integer or a double that can be converted to an integer without changing its value.

# Integer vector
vld_whole_numeric(c(1L, 2L, 3L)) # TRUE
#> [1] TRUE

# Double vector representing whole numbers
vld_whole_numeric(c(1.0, 2.0, 3.0)) # TRUE
#> [1] TRUE

# Double vector with fractional numbers
vld_whole_numeric(c(1.0, 2.2, 3.0)) # FALSE
#> [1] FALSE

The function chk_whole_number is similar to chk_whole_numeric. chk_whole_number checks if the number is of length(x) == 1L

# Integer vector
vld_whole_numeric(c(1L, 2L, 3L)) # TRUE
#> [1] TRUE
vld_whole_number(c(1L, 2L, 3L)) # FALSE
#> [1] FALSE
vld_whole_number(c(1L)) # TRUE
#> [1] TRUE

chk_count() is a special case of chk_whole_number, differing in that it ensures values are non-negative whole numbers.

# Positive integer
vld_count(1) #TRUE
#> [1] TRUE
# Zero
vld_count(0) # TRUE
#> [1] TRUE
# Negative number
vld_count(-1) # FALSE
#> [1] FALSE
# Non-whole number
vld_count(2.5) # FALSE
#> [1] FALSE

Factor Checker

Check if the function input is a factor

Function Code
chk_factor is.factor(x)
chk_character_or_factor is.character(x) || is.factor(x)

Factors can be specially confusing for users, because despite they are displayed as characters are built in top of integer vectors.

chk provides the function chk_character_or_factor() that allows detecting if the argument that the user is providing contains strings.

# Factor with specified levels

vector_fruits <- c("apple", "banana", "apple", "orange", "banana", "apple")

factor_fruits <- factor(c("apple", "banana", "apple", "orange", "banana", "apple"),
                levels = c("apple", "banana", "orange"))


is.factor(factor_fruits) # TRUE
#> [1] TRUE
vld_factor(factor_fruits) # TRUE
#> [1] TRUE

is.character(factor_fruits) # FALSE
#> [1] FALSE
vld_character(factor_fruits) # FALSE
#> [1] FALSE

vld_character_or_factor(factor_fruits) # TRUE
#> [1] TRUE

All Elements Checkers

Check if the function input has a characteristic shared by all its elements.

If you want to apply any of the previously defined functions for length(x) == 1L to the elements of a vector, you can use chk_all().

Function Code
chk_all(x, chk_fun, ...) all(vapply(x, chk_fun, TRUE, ...)) 
vld_all(c(TRUE, TRUE, FALSE), chk_lgl) # FALSE
#> [1] FALSE

Function Checker

Check if the function input is another function

formals refers to the count of the number of formal arguments

Function Code
chk_function is.function(x) && (is.null(formals) || length(formals(x)) == formals) 
vld_function(function(x) x, formals = 1) # TRUE
#> [1] TRUE
vld_function(function(x, y) x + y, formals = 1) # FALSE
#> [1] FALSE
vld_function(function(x, y) x + y, formals = 2) # TRUE
#> [1] TRUE

Name Checkers

Check if the function input has names and are valid chk_named function works with vectors, lists, data frames, and matrices that have named columns or rows. Do not confuse with check_names.

chk_valid_name function specifically designed to check if the elements of a character vector are valid R names. If you want to know what is considered a valid name, please refer to the documentation for the make.names function.

Function Code
chk_named(x) !is.null(names(x))
chk_valid_name(x) identical(make.names(x[!is.na(x)]), as.character(x[!is.na(x)])) 

vld_valid_name(c("name1", NA, "name_2", "validName"))  # TRUE
#> [1] TRUE
vld_valid_name(c(1, 2, 3))  # FALSE
#> [1] FALSE


vld_named(data.frame(a = 1:5, b = 6:10))  # TRUE
#> [1] TRUE
vld_named(list(a = 1, b = 2)) # TRUE
#> [1] TRUE
vld_named(c(a = 1, b = 2)) # TRUE 
#> [1] TRUE
vld_named(c(1, 2, 3)) # FALSE 
#> [1] FALSE

Range Checkers

Check if the function input is part of a range of values. The function input should be numeric.

Function Code
chk_range(x, range = c(0, 1)) all(x[!is.na(x)] >= range[1] & x[!is.na(x)] <= range[2])
chk_lt(x, value = 0) all(x[!is.na(x)] < value)
chk_lte(x, value = 0) all(x[!is.na(x)] <= value)
chk_gt(x, value = 0) all(x[!is.na(x)] > value)
chk_gte(x, value = 0) all(x[!is.na(x)] >= value)

Equal Checkers

Check if the function input is equal or similar to a predefined object.

The functions chk_identical(), chk_equal(), and chk_equivalent() are used to compare two objects, but they differ in how strict the comparison is.

chk_equal and chk_equivalentchecks if x and y are numerically equivalent within a specified tolerance, but chk_equivalent ignores differences in attributes.

Function Code
chk_identical(x, y) identical(x, y)
chk_equal(x, y, tolerance = sqrt(.Machine$double.eps)) vld_true(all.equal(x, y, tolerance))
chk_equivalent(x, y, tolerance = sqrt(.Machine$double.eps)) vld_true(all.equal(x, y, tolerance, check.attributes = FALSE))

In the case you want to compare the elements of a vector, you can use the check_all_* functions.

Function Code
chk_all_identical(x) length(x) < 2L || all(vapply(x, vld_identical, TRUE, y = x[[1]]))
chk_all_equal(x, tolerance = sqrt(.Machine$double.eps)) length(x) < 2L || all(vapply(x, vld_equal, TRUE, y = x[[1]], tolerance = tolerance))
chk_all_equivalent(x, tolerance = sqrt(.Machine$double.eps)) length(x) < 2L || all(vapply(x, vld_equivalent, TRUE, y = x[[1]], tolerance = tolerance)) 
vld_all_identical(c(1, 2, 3)) # FALSE
#> [1] FALSE
vld_all_identical(c(1, 1, 1)) # TRUE
#> [1] TRUE
vld_identical(c(1, 2, 3), c(1, 2, 3)) # TRUE
#> [1] TRUE

vld_all_equal(c(0.1, 0.12, 0.13))
#> [1] FALSE
vld_all_equal(c(0.1, 0.12, 0.13), tolerance = 0.2)
#> [1] TRUE
vld_equal(c(0.1, 0.12, 0.13), c(0.1, 0.12, 0.13)) # TRUE
#> [1] TRUE
vld_equal(c(0.1, 0.12, 0.13), c(0.1, 0.12, 0.4), tolerance = 0.5) # TRUE
#> [1] TRUE

x <- c(0.1, 0.1, 0.1)
y <- c(0.1, 0.12, 0.13)
attr(y, "label") <- "Numbers"
vld_equal(x, y, tolerance = 0.5) # FALSE
#> [1] FALSE
vld_equivalent(x, y, tolerance = 0.5) # TRUE
#> [1] TRUE

Order Checker

Check if the function input are numbers in increasing order chk_sorted function checks if x is sorted in non-decreasing order, ignoring any NA values.

Function Code
chk_sorted(x) !is.unsorted(x, na.rm = TRUE) 
# Checking if sorted
vld_sorted(c(1, 2, 3, NA, 4))  # TRUE
#> [1] TRUE
vld_sorted(c(3, 1, 2, NA, 4))  # FALSE
#> [1] FALSE

Set Checkers

Check if the function input is composed by certain elements

The setequal function in R is used to check if two vectors contain exactly the same elements, regardless of the order or number of repetitions.

Function Code
chk_setequal(x, values) setequal(x, values) 
vld_setequal(c(1, 2, 3), c(3, 2, 1)) # TRUE
#> [1] TRUE
vld_setequal(c(1, 2, 3), c(3, 2, 1, 4)) # FALSE
#> [1] FALSE
vld_setequal(c(1, 2, 3, 4), c(3, 2, 1)) # FALSE
#> [1] FALSE
vld_setequal(c(1, 2), c(1, 1, 1, 1, 1, 1, 2, 1)) # TRUE
#> [1] TRUE

First, the %in% function is used to check whether the elements of a vector x are present in a specified set of values. This returns a logical vector, which is then simplified by all(). The all() function checks if all values in the vector are TRUE. If the result is TRUE, it indicates that for vld_ and chk_subset(), all elements in the x vector are present in values. Similarly, for vld_ and chk_superset(), it indicates that all elements of values are present in x.

Function Code
chk_subset(x, values) all(x %in% values)
chk_not_subset(x, values) !any(x %in% values) || !length(x)
chk_superset(x, values) all(values %in% x) 

# When both function inputs have the same elements,
# all functions return TRUE

vld_setequal(c(1, 2, 3), c(3, 2, 1)) # TRUE
#> [1] TRUE
vld_subset(c(1, 2, 3), c(3, 2, 1)) # TRUE
#> [1] TRUE
vld_superset(c(1, 2, 3), c(3, 2, 1)) # TRUE
#> [1] TRUE

vld_setequal(c(1, 2), c(1, 1, 1, 1, 1, 1, 2, 1)) # TRUE
#> [1] TRUE
vld_subset(c(1, 2), c(1, 1, 1, 1, 1, 1, 2, 1)) # TRUE
#> [1] TRUE
vld_superset(c(1, 2), c(1, 1, 1, 1, 1, 1, 2, 1)) # TRUE
#> [1] TRUE

# When there are elements present in one vector but not the other,
# `vld_setequal()` will return FALSE

vld_setequal(c(1, 2, 3), c(3, 2, 1, 4)) # FALSE
#> [1] FALSE
vld_setequal(c(1, 2, 3, 4), c(3, 2, 1)) # FALSE
#> [1] FALSE

# When some elements of the `x` input are not present in `values`,
# `vld_subset()` returns FALSE
vld_subset(c(1, 2, 3, 4), c(3, 2, 1)) # FALSE
#> [1] FALSE
vld_superset(c(1, 2, 3, 4), c(3, 2, 1)) # TRUE
#> [1] TRUE

# When some elements of the `values` input are not present in `x`,
# `vld_superset()` returns FALSE

vld_subset(c(1, 2, 3), c(3, 2, 1, 4)) # TRUE
#> [1] TRUE
vld_superset(c(1, 2, 3), c(3, 2, 1, 4)) # FALSE
#> [1] FALSE

# An empty set is considered a subset of any set, and any set is a superset of an empty set.
vld_subset(c(), c("apple", "banana"))  # TRUE
#> [1] TRUE
vld_superset(c("apple", "banana"), c())  # TRUE
#> [1] TRUE

chk_orderset() validate whether a given set of values in a vector x matches a specified set of allowed values (represented by values) while preserving the order of those values.

Function Code
chk_orderset vld_equivalent(unique(x[x %in% values]), values[values %in% x]) 
vld_orderset(c("A", "B", "C"),  c("A", "B", "C", "D")) # TRUE
#> [1] TRUE
vld_orderset(c("C", "B", "A"),  c("A", "B", "C", "D")) # FALSE
#> [1] FALSE
vld_orderset(c("A", "C"),  c("A", "B", "C", "D")) # TRUE
#> [1] TRUE

Class Checkers

Check if the function input belongs to a class or type.

These functions check if x is an S3 or S4 object of the specified class.

Function Code
chk_s3_class(x, class) !isS4(x) && inherits(x, class)
chk_s4_class(x, class) isS4(x) && methods::is(x, class)

chk_is() checks if x inherits from a specified class, regardless of whether it is an S3 or S4 object.

Function Code
chk_is(x, class) inherits(x, class)

REGEX Checker

Check if the function input matches a regular expression (REGEX).

chk_match(x, regexp = ".+") checks if the regular expression pattern specified by regexp matches all the non-missing values in the vector x. If regexp it is not specified by the user, chk_match checks whether all non-missing values in x contain at least one character (regexp = “.+”)

Function Code
chk_match(x, regexp = ".+") all(grepl(regexp, x[!is.na(x)]))

Quality Checkers (Miscellaneous)

Check if the function input meet some user defined quality criteria.

chk_not_empty function checks if the length of the object is not zero. For a data frame or matrix, the length corresponds to the number of elements (not rows or columns), while for a vector or list, it corresponds to the number of elements.

chk_not_any_na function checks if there are no NA values present in the entire object.

Function Code
chk_not_empty(x) length(x) != 0L
chk_not_any_na(x) !anyNA(x) 
vld_not_empty(c()) # FALSE
#> [1] FALSE
vld_not_empty(list()) # FALSE
#> [1] FALSE
vld_not_empty(data.frame()) # FALSE
#> [1] FALSE
vld_not_empty(data.frame(a = 1:3, b = 4:6)) # TRUE
#> [1] TRUE


vld_not_any_na(data.frame(a = 1:3, b = 4:6)) # TRUE
#> [1] TRUE
vld_not_any_na(data.frame(a = c(1, NA, 3), b = c(4, 5, 6))) # FALSE
#> [1] FALSE

The chk_unique() function is designed to verify that there are no duplicates elements in a vector.

Function Code
chk_unique(x, incomparables = FALSE) !anyDuplicated(x, incomparables = incomparables) 
vld_unique(c(1, 2, 3, 4)) # TRUE
#> [1] TRUE
vld_unique(c(1, 2, 2, 4)) # FALSE
#> [1] FALSE

The function chk_length checks whether the length of x is within a specified range. It ensures that the length is at least equal to length and no more than upper. It can be used with vectors, lists and data frames.

Function Code
chk_length(x, length = 1L, upper = length) length(x) >= length && length(x) <= upper 
vld_length(c(1, 2, 3), length = 2, upper = 5)  # TRUE
#> [1] TRUE
vld_length(c("a", "b"), length = 3)  # FALSE
#> [1] FALSE

vld_length(list(a = 1, b = 2, c = 3), length = 2, upper = 4) # TRUE
#> [1] TRUE
vld_length(list(a = 1, b = 2, c = 3), length = 4) # FALSE
#> [1] FALSE

# 2 columns
vld_length(data.frame(x = 1:3, y = 4:6), length = 1, upper = 3)  # TRUE
#> [1] TRUE
vld_length(data.frame(x = 1:3, y = 4:6), length = 3)  # FALSE    
#> [1] FALSE

# length of NULL is 0
vld_length(NULL, length = 0) # TRUE
#> [1] TRUE
vld_length(NULL, length = 1) # FALSE
#> [1] FALSE

Another useful function is chk_compatible_lenghts(). This function helps to check vectors could be ‘strictly recycled’.

a <- integer(0)
b <- numeric(0)
vld_compatible_lengths(a, b) # TRUE
#> [1] TRUE

a <- 1
b <- 2
vld_compatible_lengths(a, b) # TRUE
#> [1] TRUE

a <- 1:3
b <- 1:3
vld_compatible_lengths(a, b) # TRUE
#> [1] TRUE

b <- 1
vld_compatible_lengths(a, b) # TRUE
#> [1] TRUE

b <- 1:2
vld_compatible_lengths(a, b) # FALSE
#> [1] FALSE

b <- 1:6
vld_compatible_lengths(a, b) # FALSE
#> [1] FALSE

The chk_join() function is designed to validate whether the number of rows in the resulting data frame from merging two data frames (x and y) is equal to the number of rows in the first data frame (x). This is useful when you want to ensure that a join operation does not change the number of rows in your main data frame.

Function Code
chk_join(x, y, by) identical(nrow(x), nrow(merge(x, unique(y[if (is.null(names(by))) by else names(by)]), by = by))) 
x <- data.frame(id = c(1, 2, 3), value_x = c("A", "B", "C"))
y <- data.frame(id = c(1, 2, 3), value_y = c("D", "E", "F"))
vld_join(x, y, by = "id") # TRUE
#> [1] TRUE

# Perform a join that reduces the number of rows 
y <- data.frame(id = c(1, 2, 1), value_y = c("D", "E", "F"))
vld_join(x, y, by = "id") # FALSE
#> [1] FALSE

check_ functions

The check_ functions combine several chk_ functions internally. Read the documentation for each function to learn more about its specific use.

Function Description
check_values(x, values) Checks values and S3 class of an atomic object.
check_key(x, key = character(0), na_distinct = FALSE) Checks if columns have unique rows.
check_data(x, values, exclusive, order, nrow, key) Checks column names, values, number of rows and key for a data.frame.
check_dim(x, dim, values, dim_name) Checks dimension of an object.
check_dirs(x, exists) Checks if all directories exist (or if exists = FALSE do not exist as directories or files).
check_files(x, exists) Checks if all files exist (or if exists = FALSE do not exist as files or directories).
check_names(x, names, exclusive, order) Checks the names of an object.

References

Wickham, H. (2019). Advanced R, Second Edition (2nd ed.). Chapman and Hall/CRC.