Adds a sanity check to the list of already performed sanity checks

Description

NOTE the also add_sanity_check calls this function, the parameters are documented in add_sanity_check because that function gets exported.

Usage

.add_sanity_check(
  fail_vec,
  description,
  counter_meas,
  data,
  data_name,
  example_size,
  param_name,
  call,
  fail_callback,
  .fail_vec_str,
  .generated_desc
)

Arguments

Value

see add_sanity_check

Adds a sanity check to the list of already performed sanity checks

Description

Adds a sanity check to the list of already performed sanity checks

Usage

add_sanity_check(
  fail_vec,
  description = "-",
  counter_meas = "-",
  data,
  data_name = checkmate::vname(x = data),
  example_size = 3,
  param_name = "-",
  call = h_deparsed_sys_call(which = -3),
  fail_callback
)

Arguments

Value

a list with three elements

entry_sanity_table

invisibly the sanity check that is stored internally with the other sanity checks

fail_vec

fail_vec as passed over to this function

fail

TRUE if any element of fail is TRUE. Otherwise FALSE.

All performed sanity checks can be fetched via get_sanity_checks

Examples

d <- data.frame(person_id = 1:4, bmi = c(18,23,-1,35), age = 31:34)
dummy_call <- function(x) {
  add_sanity_check(
    x$bmi < 15,
    description = "bmi above 15",
    counter_meas = "none",
    data = x,
    param_name = "bmi")
  add_sanity_check(
    x$bmi > 30,
    description = "bmi below 30",
    counter_meas = "none")
}
dummy_call(x = d)
get_sanity_checks()
add_sanity_check(
   d$bmi < 15,
   description = "bmi above 15",
   fail_callback = warning)

Removes all tracked sanity checks

Description

Removes all tracked sanity checks

Usage

clear_sanity_checks()

Returns all performed sanity checks

Description

Returns all performed sanity checks

Usage

get_sanity_checks()

Value

all sanity checks, i.e. a data.table with the following column

description

character that was provided by the user through the parameter description

additional_desc

character that provides additional information about the check that was generated by the convenience functions

data_name

name of the data set that passed to the function that performed the sanity check. This can also be specified by the user

n

a logical vector is the basis of all sanity checks. This is length of the logical vector that was used, which in general is the number of rows of the table that was checked

n_fail

how often the logical vector was TRUE

n_na

how often the logical vector was NA

counter_meas

character provided by the user about how a fail will be addressed. Note that this never happens inside a function of sanityTracker but is realized by the user after the check was performed. It is only for documentation when the results of the checks are displayed.

fail_vec_str

this captures how the actual logical vector of fails was build

param_name

usually generated by the convenience functions and it also may be a composition of more than one parameter name. However this parameter could also have been provided by the user

call

character of the call where the sanity check happend

example

if a check failed and the table is available, then some examples of rows that lead to the fail are stored here

See Also

add_sanity_check

Wrapper for add_sanity_check for internal use

Description

The convenience function usually provide some defaults like description that can be overwritten by the user through the ... argument of the convenience function. This function manages to set those values that were NOT overwritten by the user through the ... argument and then call add_sanity_check.

Usage

h_add_sanity_check(
  ellipsis,
  fail_vec,
  .generated_desc,
  data,
  data_name = "",
  param_name = "",
  call = h_deparsed_sys_call(which = -2),
  .fail_vec_str = checkmate::vname(x = fail_vec)
)

Arguments

Value

see return value of add_sanity_check

Examples

d <- data.frame(type = letters[1:4], nmb = 1:4)
# h_add_sanity_check is used on sc_col_elements()
sc_col_elements(object = d, col = "type", feasible_elements = letters[2:4])
get_sanity_checks()

Collapse a vector of characters to a string with separators

Description

Collapse a vector of characters to a string with separators

Usage

h_collapse_char_vec(v, collapse = ", ", qoute = "'")

Arguments

Value

collapsed version of v

Examples

cat(sanityTracker:::h_collapse_char_vec(v = letters[1:4]))

Extends a list with an named element if the element does not exist

Description

Extends a list with an named element if the element does not exist

Usage

h_complete_list(ell, name, value)

Arguments

Value

if ell already contained the element name, then ell is returned without being modified. Otherwise, ell is returned extended by a new element with name name and value value.

Examples

ell <- list(a = 1, b = 2)
sanityTracker:::h_complete_list(ell = ell, name = "a", value = 100)
sanityTracker:::h_complete_list(ell = ell, name = "d", value = Inf)

Simply converts a call into a character

Description

Simply converts a call into a character

Usage

h_deparsed_sys_call(which)

Arguments

Value

the call of the corresponding environment as character

Checks that the elements of a column belong to a certain set

Description

Checks that the elements of a column belong to a certain set

Usage

sc_col_elements(object, col, feasible_elements, ...)

Arguments

Value

see return object of add_sanity_check

Examples

d <- data.frame(type = letters[1:4], nmb = 1:4)
dummy_call <- function(x) {
  sc_col_elements(object = d, col = "type", feasible_elements = letters[2:4])
}
dummy_call(x = d)
get_sanity_checks()

Checks that all elements from the specified columns are in a certain range

Description

Checks that all elements from the specified columns are in a certain range

Usage

sc_cols_bounded(object, cols, rule = "(-Inf, Inf)", ...)

Arguments

Value

list of logical vectors where TRUE indicates where the check failed. Every list entry represents one of the columns specified in cols. This might be helpful if one wants to apply a counter-measure

Examples

dummy_call <- function(x) {
  sc_cols_bounded(object = iris, cols = c("Sepal.Length", "Petal.Length"),
    rule = "[1, 7.9)")
}
dummy_call(x = d)
get_sanity_checks()

Checks that all elements from the given columns are below a certain number

Description

Checks that all elements from the given columns are below a certain number

Usage

sc_cols_bounded_above(
  object,
  cols,
  upper_bound,
  include_upper_bound = TRUE,
  ...
)

Arguments

Value

list of logical vectors where TRUE indicates where the check failed. Every list entry represents one of the columns specified in cols. This might be helpful if one wants to apply a counter-measure

Checks that all elements from the given columns are above a certain number

Description

Checks that all elements from the given columns are above a certain number

Usage

sc_cols_bounded_below(
  object,
  cols,
  lower_bound,
  include_lower_bound = TRUE,
  ...
)

Arguments

Value

list of logical vectors where TRUE indicates where the check failed. Every list entry represents one of the columns specified in cols. This might be helpful if one wants to apply a counter-measure

Examples

d <- data.frame(a = c(0, 0.2, 3, Inf), b = c(1:4))
dummy_call <- function(x) {
  sc_cols_bounded_below(
    object = d, cols = c("a", "b"),
    lower_bound = 0.2,
    include_lower_bound = FALSE,
    description = "Measurements are expected to be bounded from below")
}
dummy_call(x = d)
get_sanity_checks()

Checks that all elements from the specified columns are not NA

Description

Checks that all elements from the specified columns are not NA

Usage

sc_cols_non_NA(object, cols = names(object), ..., unk_cols_callback = stop)

Arguments

Value

a list where every element is an object returned by add_sanity_check for each column specified in cols that exists in object

Examples

iris[c(1,3,5,7,9), 1] <- NA
dummy_call <- function(x) {
  sc_cols_non_NA(object = iris, description = "No NAs expected in iris")
}
dummy_call(x = iris)
get_sanity_checks()

Checks that all elements from the specified columns are positive

Description

Checks that all elements from the specified columns are positive

Usage

sc_cols_positive(object, cols, zero_feasible = TRUE, ...)

Arguments

Value

list of logical vectors where TRUE indicates where the check failed. Every list entry represents one of the columns specified in cols. This might be helpful if one wants to apply a counter-measure.

Examples

d <- data.frame(a = c(0, 0.2, 3, Inf), b = c(1:4))
dummy_call <- function(x) {
  sc_cols_positive(d, cols = c("a", "b"), zero_feasible = FALSE,
    description = "Measurements are expected to be positive")
}
dummy_call(x = d)
get_sanity_checks()

Checks that the combination of the specified columns is unique

Description

Checks that the combination of the specified columns is unique

Usage

sc_cols_unique(object, cols = names(object), ...)

Arguments

Value

see return object of add_sanity_check. Note that if a combination appears 3 times, then n_fail will increased by 3.

Examples

dummy_call <- function(x) {
  sc_cols_unique(
    object = x,
    cols = c("Species", "Sepal.Length",
             "Sepal.Width", "Petal.Length"))
}
dummy_call(x = iris)
get_sanity_checks()
get_sanity_checks()[["example"]]

Performs various checks after a left-join was performed

Description

One check is that no rows were duplicated during merge and the other check is that no columns were duplicated during merge.

Usage

sc_left_join(joined, left, right, by, ..., find_nonunique_key = TRUE)

Arguments

Value

list with two elements for the two sanity checks performed by this function. The structure of each element is as the return object of add_sanity_check.

Examples

ab <- data.table::data.table(a = 1:4, b = letters[1:4])
abc <- data.table::data.table(a = c(1:4, 2), b = letters[1:5], c = rnorm(5))
j <- merge(x = ab, y = abc, by = "a")
dummy_call <- function() {
  sc_left_join(joined = j, left = ab, right = abc, by = "a",
    description = "Left join outcome to main population")
}
dummy_call()
get_sanity_checks()