From 28c202695348b939f36f9440c9e3430ce1925477 Mon Sep 17 00:00:00 2001
From: Daniel Sjoberg <danield.sjoberg@gmail.com>
Date: Fri, 17 Sep 2021 12:03:42 -0400
Subject: [PATCH] Fixes in `tbl_summary()` and `add_p.tbl_summary()` when fct
 with all NA passed (#978)

* #977 fixes

* increment version
---
 DESCRIPTION                             |  2 +-
 NEWS.md                                 |  4 ++++
 R/utils-add_p.R                         |  2 +-
 R/utils-tbl_summary.R                   |  7 ++++++-
 tests/testthat/test-add_p.tbl_summary.R |  6 +++---
 tests/testthat/test-tbl_summary.R       | 17 +++++++++++++++++
 6 files changed, 32 insertions(+), 6 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index 3cf4b00fa2..1b27240d59 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,7 +1,7 @@
 Package: gtsummary
 Title: Presentation-Ready Data Summary and Analytic Result
     Tables
-Version: 1.4.2.9008
+Version: 1.4.2.9009
 Authors@R: 
     c(person(given = "Daniel D.",
              family = "Sjoberg",
diff --git a/NEWS.md b/NEWS.md
index b7ecbd6fd4..0d8b71aaee 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,5 +1,9 @@
 # gtsummary (development version)
 
+* Fix in `tbl_summary()` when a factor variable is passed that is all NA with no specified levels. (#977)
+
+* Fix in `add_p.tbl_summary()` when a factor variable with all NA values is passed. (#977)
+
 * New function `modify_cols_merge()` that can merge two or more columns in a gtsummary table. (#939)
 
 * Added new function `add_prop_ci()` that adds a new column with the confidence interval for proportions reported in `tbl_summary()`. Cannot be used if `tbl_summary()` included a `by=` argument. (#868)
diff --git a/R/utils-add_p.R b/R/utils-add_p.R
index ec8cb45ca0..d2f978d8bd 100644
--- a/R/utils-add_p.R
+++ b/R/utils-add_p.R
@@ -241,7 +241,7 @@
     )
 
   # if expected counts >= 5 for all cells, chisq, otherwise Fishers exact
-  if (min_exp >= 5) {
+  if (isTRUE(min_exp >= 5 || is.nan(min_exp))) {
     test_func <-
       get_theme_element("add_p.tbl_summary-attr:test.categorical") %||%
       getOption("gtsummary.add_p.test.categorical", default = "chisq.test.no.correct")
diff --git a/R/utils-tbl_summary.R b/R/utils-tbl_summary.R
index 5ce9e24e2b..bab8872466 100644
--- a/R/utils-tbl_summary.R
+++ b/R/utils-tbl_summary.R
@@ -204,9 +204,14 @@ assign_summary_type <- function(data, variable, summary_type, value,
           if (inherits(data[[variable]], "character")) {
             return("dichotomous")
           }
-          if (inherits(data[[variable]], "factor")) {
+          if (inherits(data[[variable]], "factor") &&
+              !rlang::is_empty(attr(data[[variable]], "levels"))) {
             return("categorical")
           }
+          if (inherits(data[[variable]], "factor") &&
+              rlang::is_empty(attr(data[[variable]], "levels"))) {
+            return("dichotomous")
+          }
         }
 
         # numeric variables that are 0 and 1 only, will be dichotomous
diff --git a/tests/testthat/test-add_p.tbl_summary.R b/tests/testthat/test-add_p.tbl_summary.R
index d7f8486561..c9f378ab07 100644
--- a/tests/testthat/test-add_p.tbl_summary.R
+++ b/tests/testthat/test-add_p.tbl_summary.R
@@ -398,13 +398,13 @@ test_that("no error with missing data", {
   expect_message(
     t1 <-
       mtcars %>%
-      mutate(mpg = NA, hp = NA) %>%
-      select(mpg, hp, am) %>%
+      mutate(mpg = NA, hp = NA, has_banana = factor(NA, levels = c("Yes", "No"))) %>%
+      select(has_banana, mpg, hp, am) %>%
       tbl_summary(by = "am", type = hp ~ "continuous") %>%
       add_p()
   )
   expect_equal(
     t1 %>% as_tibble(col_labels = FALSE) %>% dplyr::pull(p.value),
-    rep_len(NA_character_, 4)
+    rep_len(NA_character_, 8)
   )
 })
diff --git a/tests/testthat/test-tbl_summary.R b/tests/testthat/test-tbl_summary.R
index c307008d44..82b586c2c1 100644
--- a/tests/testthat/test-tbl_summary.R
+++ b/tests/testthat/test-tbl_summary.R
@@ -643,3 +643,20 @@ test_that("no error when by variable omitted from include", {
     NA
   )
 })
+
+test_that("no error with factor variable with all NA and no specifed levels", {
+  expect_error(
+    tbl <-
+      trial %>%
+      dplyr::mutate(
+        has_banana = factor(NA) # We don't know which patients have a banana
+      ) %>%
+      tbl_summary(by = trt, include = has_banana) %>%
+      as_tibble(col_labels = FALSE),
+    NA
+  )
+  expect_equal(
+    tbl$stat_1,
+    c("0 (NA%)", "98")
+  )
+})