assert_non_na()

See #21

NAMESPACE

@@ -1,5 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export(assert_non_na)
 export(clump_month_date)
 export(column_class_headstart)
 export(column_rename_headstart)

R/assert.R

@@ -0,0 +1,42 @@
+#' @name assert
+#' @title Assert vector characteristics
+#'  
+#' @description Assert a vector meets important data-quality requirements.
+#' 
+#' @param x Vector to inspect.  Required.
+#' @param class_vector The required [base::class()] of the vector.  If the parameter is missing, the object's class is not checked.
+#' @param proportion_minimum The (inclusive) minimum proportion of the vector's elements that should meet the requirement. If missing, all elements must pass.
+#' 
+#' @examples 
+#' requireNamespace("OuhscMunge") 
+#' OuhscMunge::assert_non_na(1:30, "integer")
+#' \dontrun{
+#' OuhscMunge::assert_non_na(c(1:30, NA_integer_), "integer")
+#' }
+
+
+#' @export
+assert_non_na <- function( x, class_vector, proportion_minimum ) {
+
+  if( !missing(class_vector) & !inherits(x, class_vector) ) {
+    stop("The vector must inherit from the class `", class_vector, "`, but it is a `", class(x), "`.")
+  }
+
+
+  if( missing(proportion_minimum) ) {
+
+    missing_count <- sum(is.na(x))
+    if( missing_count > 0L ) {
+      stop("The vector should not have any NA values, but `", missing_count, "` elements were NA.")
+    }
+
+  } else {
+
+    proportion_actual <- mean(!is.na(x))
+    if( proportion_actual <  proportion_minimum) {
+      stop("The vector must have a proportion of at least `", proportion_minimum, "` of nonmissing elements.  However the actual nonmissing proportion is `", proportion_actual, "1.")
+    }
+
+  }
+
+}

tests/testthat/test-assert.R

@@ -0,0 +1,50 @@
+library(testthat)
+
+###########
+context("Assert")
+###########
+
+test_that("assert_non_na -all - passes", {
+  expect_silent( assert_non_na(1:100                                                              , "integer"   ))
+  expect_silent( assert_non_na(runif(100)                                                         , "numeric"   ))
+  expect_silent( assert_non_na(letters                                                            , "character" ))
+  expect_silent( assert_non_na(sample(c(T,F), size=100, replace = TRUE)                           , "logical"   ))
+  expect_silent( assert_non_na(seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days")  , "Date"      ))
+
+  expect_silent( assert_non_na(1:100                                                              , "integer"   , 1.0))
+  expect_silent( assert_non_na(runif(100)                                                         , "numeric"   , 1.0))
+  expect_silent( assert_non_na(letters                                                            , "character" , 1.0))
+  expect_silent( assert_non_na(sample(c(T,F), size=100, replace = TRUE)                           , "logical"   , 1.0))
+  expect_silent( assert_non_na(seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days")  , "Date"      , 1.0))
+})
+
+test_that("assert_non_na -all - fails", {
+  expect_error( assert_non_na(c(NA, 1:100                                                                       ) , "integer"   ))
+  expect_error( assert_non_na(c(NA, runif(100)                                                                  ) , "numeric"   ))
+  expect_error( assert_non_na(c(NA, letters                                                                     ) , "character" ))
+  expect_error( assert_non_na(c(NA, sample(c(T,F), size=100, replace = TRUE)                                    ) , "logical"   ))
+  expect_error( assert_non_na(c(as.Date(NA), seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days")  ) , "Date"      ))
+})
+
+
+test_that("assert_non_na -proportion - passes", {
+  expect_silent( assert_non_na(1:1                                                               , "integer"   , 1.0))
+  expect_silent( assert_non_na(runif(100)                                                        , "numeric"   , 1.0))
+  expect_silent( assert_non_na(letters                                                           , "character" , 1.0))
+  expect_silent( assert_non_na(sample(c(T,F), size=100, replace = TRUE)                          , "logical"   , 1.0))
+  expect_silent( assert_non_na(seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days") , "Date"      , 1.0))
+
+  expect_silent( assert_non_na(c(NA, 1:100                                                                        ) , "integer"   , 0.5))
+  expect_silent( assert_non_na(c(NA, runif(100)                                                                   ) , "numeric"   , 0.5))
+  expect_silent( assert_non_na(c(NA, letters                                                                      ) , "character" , 0.5))
+  expect_silent( assert_non_na(c(NA, sample(c(T,F), size=100, replace = TRUE)                                     ) , "logical"   , 0.5))
+  expect_silent( assert_non_na(c(as.Date(NA), seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days")   ) , "Date"      , 0.5))
+})
+
+test_that("assert_non_na -proportion - fails", {
+  expect_error( assert_non_na(c(NA, 1:100                                                                       ) , "integer"   , 0.999))
+  expect_error( assert_non_na(c(NA, runif(100)                                                                  ) , "numeric"   , 0.999))
+  expect_error( assert_non_na(c(NA, letters                                                                     ) , "character" , 0.999))
+  expect_error( assert_non_na(c(NA, sample(c(T,F), size=100, replace = TRUE)                                    ) , "logical"   , 0.999))
+  expect_error( assert_non_na(c(as.Date(NA), seq.Date(as.Date("2015-01-02"), as.Date("2017-04-30"), by="days")  ) , "Date"      , 0.999))
+})