test-derive_merged.R
In admiral: ADaM in R Asset Library

adsl <- tibble::tribble(
  ~USUBJID, ~SEX, ~COUNTRY,
  "ST42-1", "F",  "AUT",
  "ST42-2", "M",  "MWI",
  "ST42-3", "M",  "NOR",
  "ST42-4", "F",  "UGA"
) %>% mutate(STUDYID = "ST42")

adsl1 <- tibble::tribble(
  ~ID, ~SEX, ~COUNTRY,
  "ST42-1", "F", "AUT",
  "ST42-2", "M", "MWI",
  "ST42-3", "M", "NOR",
  "ST42-4", "F", "UGA"
) %>% mutate(STUDYID = "ST42")

adsl2 <- tibble::tribble(
  ~ID, ~SEX, ~COUNTRY,
  "ST42-1", "F", "AUT",
  "ST42-1", "F", "NOR",
  "ST42-2", "M", "MWI",
  "ST42-3", "M", "NOR",
  "ST42-4", "F", "UGA"
) %>% mutate(STUDYID = "ST42")


advs <- tibble::tribble(
  ~USUBJID, ~PARAMCD, ~AVISIT,    ~AVAL,
  "ST42-1", "WEIGHT", "BASELINE", 66,
  "ST42-1", "WEIGHT", "Week 2",   68,
  "ST42-2", "WEIGHT", "BASELINE", 88,
  "ST42-3", "WEIGHT", "Week 2",   55,
  "ST42-3", "WEIGHT", "Week 4",   50
) %>% mutate(STUDYID = "ST42")

advs1 <- tibble::tribble(
  ~ID, ~PARAMCD, ~AVISIT, ~AVAL,
  "ST42-1", "WEIGHT", "BASELINE", 66,
  "ST42-1", "WEIGHT", "Week 2", 68,
  "ST42-2", "WEIGHT", "BASELINE", 88,
  "ST42-3", "WEIGHT", "Week 2", 55,
  "ST42-3", "WEIGHT", "Week 4", 50
) %>% mutate(STUDYID = "ST42")


ex <- tibble::tribble(
  ~USUBJID, ~EXSTDTC,
  "ST42-1", "2020-12-07",
  "ST42-1", "2020-12-14",
  "ST42-2", "2021-01-12T12:00:00",
  "ST42-2", "2021-01-26T13:21",
  "ST42-3", "2021-03-02"
) %>% mutate(STUDYID = "ST42")

vs <- tibble::tribble(
  ~USUBJID, ~VSTESTCD, ~VSTEST, ~VSORRES, ~VSSEQ,
  "ST42-1", "DIABP", "Diastolic Blood Pressure", 64, 1,
  "ST42-1", "DIABP", "Diastolic Blood Pressure", 83, 2,
  "ST42-1", "WEIGHT", "Weight", 120, 3,
  "ST42-2", "WEIGHT", "Weight", 110, 1,
  "ST42-2", "HEIGHT", "Height", 58, 2
) %>% mutate(STUDYID = "ST42")

# derive_vars_merged ----
## Test 1: merge all variables ----
test_that("derive_vars_merged Test 1: merge all variables", {
  actual <- derive_vars_merged(advs,
    dataset_add = adsl,
    by_vars = exprs(STUDYID, USUBJID)
  )

  expected <- left_join(advs, adsl, by = c("STUDYID", "USUBJID"))

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "AVISIT")
  )
})

## Test 2: merge selected variables ----
test_that("derive_vars_merged Test 2: merge selected variables", {
  actual <- derive_vars_merged(advs,
    dataset_add = adsl,
    by_vars = exprs(USUBJID),
    new_vars = exprs(SEX)
  )

  expected <- left_join(advs, select(adsl, USUBJID, SEX), by = "USUBJID")

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "AVISIT")
  )
})

## Test 3: merge last value and flag matched by groups ----
test_that("derive_vars_merged Test 3: merge last value and flag matched by groups", {
  actual <- derive_vars_merged(adsl,
    dataset_add = advs,
    order = exprs(AVAL),
    by_vars = exprs(STUDYID, USUBJID),
    new_vars = exprs(WEIGHTBL = AVAL),
    mode = "last",
    match_flag = matched
  )
  expected <- adsl %>% mutate(
    WEIGHTBL = c(68, 88, 55, NA),
    matched = c(TRUE, TRUE, TRUE, NA)
  )

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID")
  )
})

## Test 4: error if variable in both datasets ----
test_that("derive_vars_merged Test 4: error if variable in both datasets", {
  expect_error(
    derive_vars_merged(advs,
      dataset_add = adsl,
      by_vars = exprs(USUBJID)
    ),
    regexp = ""
  )
})

## Test 5: by_vars with rename ----
test_that("derive_vars_merged Test 5: by_vars with rename", {
  actual <- derive_vars_merged(advs,
    dataset_add = adsl1,
    by_vars = exprs(STUDYID, USUBJID = ID),
    filter_add = SEX == "F"
  )

  adsl_1 <- adsl1 %>% filter(SEX == "F")
  expected <- left_join(advs, adsl_1, by = c("STUDYID", "USUBJID" = "ID"))

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "AVISIT")
  )
})

## Test 6: expressions for new_vars and missing_values ----
test_that("derive_vars_merged Test 6: expressions for new_vars and missing_values", {
  actual <- derive_vars_merged(
    adsl,
    dataset_add = advs,
    by_vars = exprs(USUBJID),
    order = exprs(AVISIT),
    new_vars = exprs(LASTVIS = str_to_upper(AVISIT)),
    mode = "last",
    missing_values = exprs(LASTVIS = "UNKNOWN")
  )

  expected <-
    mutate(adsl, LASTVIS = c("WEEK 2", "BASELINE", "WEEK 4", "UNKNOWN"))


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = "USUBJID"
  )
})

## Test 7: use new variables in filter_add and order ----
test_that("derive_vars_merged Test 7: use new variables in filter_add and order", {
  expected <- tibble::tribble(
    ~USUBJID, ~TRTSDT,      ~TRTSSEQ,
    "ST42-1", "2020-12-14",        2,
    "ST42-2", "2021-01-26",        2,
    "ST42-3", NA,                 NA,
    "ST42-4", NA,                 NA
  ) %>% mutate(
    STUDYID = "ST42",
    TRTSDT = ymd(TRTSDT)
  )

  ex <- tibble::tribble(
    ~USUBJID, ~EXSTDTC,              ~EXSEQ,
    "ST42-1", "2020-12-07",               1,
    "ST42-1", "2020-12-14",               2,
    "ST42-2", "2021-01-12T12:00:00",      1,
    "ST42-2", "2021-01-26T13:21",         2,
    "ST42-3", "2021-03",                  1
  ) %>% mutate(STUDYID = "ST42")

  actual <- derive_vars_merged(
    select(adsl, STUDYID, USUBJID),
    dataset_add = ex,
    by_vars = exprs(USUBJID),
    order = exprs(TRTSDT),
    new_vars = exprs(TRTSDT = convert_dtc_to_dt(EXSTDTC), TRTSSEQ = EXSEQ),
    filter_add = !is.na(TRTSDT),
    mode = "last"
  )

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = "USUBJID"
  )
})

## Test 8: warning if not unique w.r.t the by variables and the order ----
test_that("derive_vars_merged Test 8: warning if not unique w.r.t the by variables and the order", {
  expect_warning(
    actual <- derive_vars_merged(advs,
      dataset_add = adsl2,
      by_vars = exprs(STUDYID, USUBJID = ID),
      order = exprs(ID),
      mode = "last",
      check_type = "warning"
    ),
    regexp = ""
  )
})

## Test 9: error if not unique w.r.t the by variables and the order ----
test_that("derive_vars_merged Test 9: error if not unique w.r.t the by variables and the order", {
  expect_error(
    actual <- derive_vars_merged(advs,
      dataset_add = adsl2,
      by_vars = exprs(STUDYID, USUBJID = ID),
      order = exprs(ID),
      mode = "last",
      check_type = "error",
      duplicate_msg = "Duplicate records present!"
    ),
    regexp = ""
  )
})

## Test 10: error if variables in missing_values but not in new_vars ----
test_that("derive_vars_merged Test 10: error if variables in missing_values but not in new_vars", {
  expect_error(
    derive_vars_merged(
      adsl,
      dataset_add = advs,
      by_vars = exprs(USUBJID),
      order = exprs(AVISIT),
      new_vars = exprs(LASTVIS = str_to_upper(AVISIT)),
      mode = "last",
      missing_values = exprs(LASTVIS = "UNKNOWN", LASTVISN = -1)
    ),
    regexp = "The variables `LASTVISN` were specified for `missing_values` but not for `new_vars`.",
    fixed = TRUE
  )
})

# derive_var_merged_cat ----

## Test 11: deprecation error ----
test_that("derive_var_merged_cat Test 11: deprecation error", {
  get_vscat <- function(x) {
    if_else(x == "BASELINE", "BASELINE", "POST-BASELINE")
  }

  expect_error(
    derive_var_merged_cat(
      adsl,
      dataset_add = advs,
      by_vars = exprs(USUBJID),
      new_var = LSTVSCAT,
      source_var = AVISIT,
      cat_fun = get_vscat,
      order = exprs(AVISIT),
      mode = "last",
      missing_value = "MISSING"
    ),
    class = "lifecycle_error_deprecated"
  )
})


# derive_var_merged_exist_flag ----
## Test 12: merge existence flag ----
test_that("derive_var_merged_exist_flag Test 12: merge existence flag", {
  actual <- derive_var_merged_exist_flag(
    adsl,
    dataset_add = advs,
    by_vars = exprs(USUBJID),
    new_var = VSEVALFL,
    condition = AVISIT == "BASELINE"
  )

  expected <-
    mutate(adsl, VSEVALFL = c("Y", "Y", NA_character_, NA_character_))


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = "USUBJID"
  )
})

## Test 13: by_vars with rename ----
test_that("derive_var_merged_exist_flag Test 13: by_vars with rename", {
  actual <- derive_var_merged_exist_flag(
    adsl,
    dataset_add = advs1,
    by_vars = exprs(USUBJID = ID),
    new_var = VSEVALFL,
    condition = AVISIT == "BASELINE",
  )

  expected <-
    mutate(adsl, VSEVALFL = c("Y", "Y", NA_character_, NA_character_))


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = "USUBJID"
  )
})

# derive_var_merged_character ----

## Test 14: deprecation error ----
test_that("derive_var_merged_character Test 14: deprecation error", {
  expect_error(
    derive_var_merged_character(
      adsl,
      dataset_add = advs,
      by_vars = exprs(USUBJID),
      order = exprs(AVISIT),
      new_var = LASTVIS,
      source_var = AVISIT,
      mode = "last"
    ),
    class = "lifecycle_error_deprecated"
  )
})


# derive_vars_merged_lookup ----
## Test 15: merge lookup table ----
test_that("derive_vars_merged_lookup Test 15: merge lookup table", {
  param_lookup <- tibble::tribble(
    ~VSTESTCD, ~VSTEST, ~PARAMCD, ~DESCRIPTION,
    "WEIGHT", "Weight", "WEIGHT", "Weight (kg)",
    "HEIGHT", "Height", "HEIGHT", "Height (cm)",
    "BMI", "Body Mass Index", "BMI", "Body Mass Index(kg/m^2)"
  )

  attr(param_lookup$VSTESTCD, "label") <- "Vital Signs Test Short Name"
  attr(param_lookup$VSTEST, "label") <- "Vital Signs Test Name"

  expect_message(
    actual <- derive_vars_merged_lookup(
      vs,
      dataset_add = param_lookup,
      by_vars = exprs(VSTESTCD, VSTEST),
      new_var = exprs(PARAMCD, PARAM = DESCRIPTION),
      print_not_mapped = TRUE
    ),
    regex = "^List of `VSTESTCD` and `VSTEST` not mapped: .*$"
  )

  expected <-
    left_join(vs, param_lookup, by = c("VSTESTCD", "VSTEST")) %>%
    rename(PARAM = DESCRIPTION)


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "VSSEQ", "VSTESTCD")
  )
})



## the lookup table
## Test 16:  all by_vars have records in the lookup table ----
test_that("derive_vars_merged_lookup Test 16:  all by_vars have records in the lookup table", {
  param_lookup <- tibble::tribble(
    ~VSTESTCD, ~VSTEST, ~PARAMCD, ~DESCRIPTION,
    "WEIGHT", "Weight", "WEIGHT", "Weight (kg)",
    "HEIGHT", "Height", "HEIGHT", "Height (cm)",
    "BMI", "Body Mass Index", "BMI", "Body Mass Index(kg/m^2)",
    "DIABP", "Diastolic Blood Pressure", "DIABP", "Diastolic Blood Pressure (mmHg)"
  )

  attr(param_lookup$VSTESTCD, "label") <- "Vital Signs Test Short Name"
  attr(param_lookup$VSTEST, "label") <- "Vital Signs Test Name"

  expect_message(
    actual <- derive_vars_merged_lookup(
      vs,
      dataset_add = param_lookup,
      by_vars = exprs(VSTESTCD, VSTEST),
      new_var = exprs(PARAMCD, PARAM = DESCRIPTION),
      print_not_mapped = TRUE
    ),
    regex = "All `VSTESTCD` and `VSTEST` are mapped."
  )

  expected <-
    left_join(vs, param_lookup, by = c("VSTESTCD", "VSTEST")) %>%
    rename(PARAM = DESCRIPTION)


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "VSSEQ", "VSTESTCD")
  )
})

## Test 17: by_vars with rename ----
test_that("derive_vars_merged_lookup Test 17: by_vars with rename", {
  param_lookup <- tibble::tribble(
    ~TESTCD, ~VSTEST, ~PARAMCD, ~DESCRIPTION,
    "WEIGHT", "Weight", "WEIGHT", "Weight (kg)",
    "HEIGHT", "Height", "HEIGHT", "Height (cm)",
    "BMI", "Body Mass Index", "BMI", "Body Mass Index(kg/m^2)"
  )

  attr(param_lookup$TESTCD, "label") <- "Vital Signs Test Short Name"
  attr(param_lookup$VSTEST, "label") <- "Vital Signs Test Name"

  expect_message(
    actual <- derive_vars_merged_lookup(
      vs,
      dataset_add = param_lookup,
      by_vars = exprs(VSTESTCD = TESTCD, VSTEST),
      new_var = exprs(PARAMCD, PARAM = DESCRIPTION),
      print_not_mapped = TRUE
    ),
    regex = "^List of `VSTESTCD` and `VSTEST` not mapped: .*$"
  )

  expected <-
    left_join(vs, param_lookup, by = c("VSTESTCD" = "TESTCD", "VSTEST")) %>%
    rename(PARAM = DESCRIPTION)


  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("USUBJID", "VSSEQ", "VSTESTCD")
  )
})


# get_not_mapped ----
## Test 18: not all by_vars have records in the lookup table ----
test_that("get_not_mapped Test 18: not all by_vars have records in the lookup table", {
  param_lookup <- tibble::tribble(
    ~VSTESTCD, ~VSTEST, ~PARAMCD, ~DESCRIPTION,
    "WEIGHT", "Weight", "WEIGHT", "Weight (kg)",
    "HEIGHT", "Height", "HEIGHT", "Height (cm)",
    "BMI", "Body Mass Index", "BMI", "Body Mass Index(kg/m^2)"
  )

  attr(param_lookup$VSTESTCD, "label") <- "Vital Signs Test Short Name"
  attr(param_lookup$VSTEST, "label") <- "Vital Signs Test Name"

  expect_message(
    act_vs_param <- derive_vars_merged_lookup(
      vs,
      dataset_add = param_lookup,
      by_vars = exprs(VSTESTCD, VSTEST),
      new_var = exprs(PARAMCD, PARAM = DESCRIPTION),
      print_not_mapped = TRUE
    ),
    regex = "^List of `VSTESTCD` and `VSTEST` not mapped: .*$"
  )

  actual <- get_not_mapped()

  expected <- left_join(vs, param_lookup, by = c("VSTESTCD", "VSTEST")) %>%
    rename(PARAM = DESCRIPTION) %>%
    filter(is.na(PARAMCD)) %>%
    select(VSTESTCD, VSTEST) %>%
    distinct()

  expect_dfs_equal(
    base = expected,
    compare = actual,
    keys = c("VSTESTCD", "VSTEST")
  )
})

# derive_var_merged_summary ----
## Test 19: dataset == dataset_add, no filter ----
test_that("derive_var_merged_summary Test 19: dataset == dataset_add, no filter", {
  expected <- tibble::tribble(
    ~AVISIT,  ~ASEQ, ~AVAL, ~MEANVIS,
    "WEEK 1",     1,    10,       10,
    "WEEK 1",     2,    NA,       10,
    "WEEK 2",     3,    NA,       NA,
    "WEEK 3",     4,    42,       42,
    "WEEK 4",     5,    12,       13,
    "WEEK 4",     6,    12,       13,
    "WEEK 4",     7,    15,       13
  )

  adbds <- select(expected, -MEANVIS)

  expect_dfs_equal(
    base = expected,
    compare = derive_var_merged_summary(
      adbds,
      dataset_add = adbds,
      by_vars = exprs(AVISIT),
      new_var = MEANVIS,
      analysis_var = AVAL,
      summary_fun = function(x) mean(x, na.rm = TRUE)
    ),
    keys = c("AVISIT", "ASEQ")
  )
})

## Test 20: dataset != dataset_add, filter ----
test_that("derive_var_merged_summary Test 20: dataset != dataset_add, filter", {
  expected <- tibble::tribble(
    ~USUBJID, ~MEANPBL,
    "1",          13.5,
    "2",            NA,
    "3",          42.0
  )

  adbds <- tibble::tribble(
    ~USUBJID, ~ADY, ~AVAL,
    "1",        -3,    10,
    "1",         2,    12,
    "1",         8,    15,
    "3",         4,    42
  )

  adsl <- select(expected, -MEANPBL)

  expect_dfs_equal(
    base = expected,
    compare = derive_var_merged_summary(
      adsl,
      dataset_add = adbds,
      by_vars = exprs(USUBJID),
      new_var = MEANPBL,
      filter_add = ADY > 0,
      analysis_var = AVAL,
      summary_fun = function(x) mean(x, na.rm = TRUE)
    ),
    keys = c("USUBJID")
  )
})

## Test 21: by_vars with rename ----
test_that("derive_var_merged_summary Test 21: by_vars with rename", {
  expected <- tibble::tribble(
    ~AVISIT,  ~ASEQ, ~AVAL, ~MEANVIS,
    "WEEK 1",     1,    10,       10,
    "WEEK 1",     2,    NA,       10,
    "WEEK 2",     3,    NA,       NA,
    "WEEK 3",     4,    42,       42,
    "WEEK 4",     5,    12,       13,
    "WEEK 4",     6,    12,       13,
    "WEEK 4",     7,    15,       13
  )
  adbds <- select(expected, -MEANVIS)
  adbds1 <- select(expected, -MEANVIS) %>%
    rename(VISIT = AVISIT)

  expect_dfs_equal(
    base = expected,
    compare = derive_var_merged_summary(
      adbds,
      dataset_add = adbds1,
      by_vars = exprs(AVISIT = VISIT),
      new_var = MEANVIS,
      analysis_var = AVAL,
      summary_fun = function(x) mean(x, na.rm = TRUE)
    ),
    keys = c("AVISIT", "ASEQ")
  )
})