issue #633

R/Utilities.R

@@ -872,6 +872,20 @@ TADA_ConvertSpecialChars <- function(
     clean.data <- TADA_OrderCols(clean.data)
   }
 
+  # Flag result values that do not have an associated result unit
+  if (
+    col %in%
+      c("ResultMeasureValue", "TADA.ResultMeasureValue") &&
+      "TADA.ResultMeasure.MeasureUnitCode" %in% names(clean.data)
+  ) {
+    clean.data[[flagcol]] <- ifelse(
+      is.na(clean.data$TADA.ResultMeasure.MeasureUnitCode) |
+        trimws(clean.data$TADA.ResultMeasure.MeasureUnitCode) == "",
+      "No unit associated with result value",
+      clean.data[[flagcol]]
+    )
+  }
+
   if (flaggedonly == FALSE) {
     if (clean == TRUE) {
       clean.data <- clean.data |>
@@ -886,6 +900,20 @@ TADA_ConvertSpecialChars <- function(
             )
         )
 
+      # Remove records with missing result units when cleaning result values
+      if (
+        col %in%
+          c("ResultMeasureValue", "TADA.ResultMeasureValue") &&
+          "TADA.ResultMeasure.MeasureUnitCode" %in% names(clean.data)
+      ) {
+        clean.data <- clean.data |>
+          dplyr::filter(
+            !is.na(TADA.ResultMeasure.MeasureUnitCode),
+            # trimws incorporates " " into this as well as ""
+            trimws(TADA.ResultMeasure.MeasureUnitCode) != ""
+          )
+      }
+
       return(clean.data)
     }
 
@@ -903,6 +931,7 @@ TADA_ConvertSpecialChars <- function(
             "Text",
             "Non-ASCII Character(s)",
             "Result Value/Unit Cannot Be Estimated From Detection Limit",
+            "No unit associated with result value",
             "Coerced to NA"
           )
       )

tests/testthat/test-Utilities.R

@@ -38,6 +38,27 @@ test_that("Column names do not contain the pattern 'TADA.TADA.'", {
   )
 })
 
+test_that("TADA_ConvertSpecialChars removes missing result units when clean is TRUE", {
+  testdat <- TADA_RandomTestingData()
+
+  # Ensure the test includes both NA and blank result units
+  testdat$TADA.ResultMeasure.MeasureUnitCode[1] <- NA_character_
+  testdat$TADA.ResultMeasure.MeasureUnitCode[2] <- ""
+
+  result <- TADA_ConvertSpecialChars(
+    testdat,
+    col = "TADA.ResultMeasureValue",
+    clean = TRUE
+  )
+
+  expect_true(is.numeric(result$TADA.ResultMeasureValue))
+  expect_false(any(is.na(result$TADA.ResultMeasureValue)))
+
+  expect_false(any(is.na(result$TADA.ResultMeasure.MeasureUnitCode)))
+
+  expect_false(any(result$TADA.ResultMeasure.MeasureUnitCode == ""))
+})
+
 test_that("Column names do not contain the pattern 'TADA.TADA.'", {
   test_TADA.TADA. <- TADA_AutoClean(Data_R5_TADAPackageDemo)
   # Create a logical vector indicating which columns contain the pattern
@@ -50,6 +71,106 @@ test_that("Column names do not contain the pattern 'TADA.TADA.'", {
   )
 })
 
+test_that("TADA_ConvertSpecialChars removes rows with missing result units when clean = TRUE", {
+  testdat <- Data_Nutrients_UT[1:4, ]
+
+  testdat$ResultMeasureValue <- c("1.2", "2.3", "3.4", "4.5")
+  testdat$ResultMeasure.MeasureUnitCode <- c("mg/L", NA_character_, "", "ug/L")
+
+  result <- TADA_ConvertSpecialChars(
+    testdat,
+    col = "ResultMeasureValue",
+    clean = TRUE
+  )
+
+  # Rows with NA or blank result units should be removed
+  expect_equal(nrow(result), 2)
+
+  # Confirm that the correct converted values remain
+  expect_equal(result$TADA.ResultMeasureValue, c(1.2, 4.5))
+
+  expect_equal(result$TADA.ResultMeasure.MeasureUnitCode, c("MG/L", "UG/L"))
+
+  # Confirm the output columns have the expected types and values
+  expect_true(is.numeric(result$TADA.ResultMeasureValue))
+  expect_false(any(is.na(result$TADA.ResultMeasureValue)))
+
+  expect_false(any(is.na(result$TADA.ResultMeasure.MeasureUnitCode)))
+
+  expect_false(any(result$TADA.ResultMeasure.MeasureUnitCode == ""))
+})
+
+test_that("TADA_ConvertSpecialChars flags rows with missing result units when clean = FALSE", {
+  testdat <- Data_Nutrients_UT[1:5, ]
+
+  testdat$ResultMeasureValue <- c("1.2", "2.3", "3.4", "4.5", "5.6")
+  testdat$ResultMeasure.MeasureUnitCode <- c(
+    "mg/L",
+    NA_character_,
+    "",
+    "   ",
+    "ug/L"
+  )
+
+  result <- TADA_ConvertSpecialChars(
+    testdat,
+    col = "ResultMeasureValue",
+    clean = FALSE
+  )
+
+  # No rows should be removed
+  expect_equal(nrow(result), 5)
+
+  # Confirm all result values are converted and retained
+  expect_equal(result$TADA.ResultMeasureValue, c(1.2, 2.3, 3.4, 4.5, 5.6))
+
+  # Confirm missing, blank, and whitespace-only units receive the new flag
+  expect_equal(
+    result$TADA.ResultMeasureValueDataTypes.Flag,
+    c(
+      "Numeric",
+      "No unit associated with result value",
+      "No unit associated with result value",
+      "No unit associated with result value",
+      "Numeric"
+    )
+  )
+
+  # Confirm result values remain numeric
+  expect_true(is.numeric(result$TADA.ResultMeasureValue))
+  expect_false(any(is.na(result$TADA.ResultMeasureValue)))
+})
+
+test_that("TADA_ConvertSpecialChars returns missing-unit rows when flaggedonly = TRUE", {
+  testdat <- Data_Nutrients_UT[1:5, ]
+
+  testdat$ResultMeasureValue <- c("1.2", "2.3", "3.4", "4.5", "5.6")
+  testdat$ResultMeasure.MeasureUnitCode <- c(
+    "mg/L",
+    NA_character_,
+    "",
+    "   ",
+    "ug/L"
+  )
+
+  result <- TADA_ConvertSpecialChars(
+    testdat,
+    col = "ResultMeasureValue",
+    flaggedonly = TRUE
+  )
+
+  # Only rows with missing, blank, or whitespace-only units should remain
+  expect_equal(nrow(result), 3)
+
+  # Confirm the expected result values are returned
+  expect_equal(result$TADA.ResultMeasureValue, c(2.3, 3.4, 4.5))
+
+  # Confirm all returned rows have the missing-unit flag
+  expect_true(all(
+    result$TADA.ResultMeasureValueDataTypes.Flag ==
+      "No unit associated with result value"
+  ))
+})
 
 test_that("Only numeric data remains after running TADA_ConvertSpecialChars clean = TRUE", {
   testdat <- TADA_RandomTestingData(