Ensure that tz is respected in sas_numeric_to_date(); test update…

… for recent change; clarify minor breaking change for `sas_numeric_to_date()`

NEWS.md

@@ -8,6 +8,7 @@ These are all minor breaking changes resulting from enhancements and are not exp
 
 * When `tabyl()` is called on a data.frame containing labels, it now displays the label attribute as the name of the first column in the the resulting `tabyl` object (@olivroy, #394). This may break subsequent code that refers to the output of such a `tabyl` by column name. To maintain the previous behavior of ignoring variable labels, you can remove the labels with a function like `haven::zap_labels()` or `labelled::remove_labels()` before calling `tabyl()`.
 
+* `sas_numeric_to_date()` now warns for timezones other than "UTC" due to the way that SAS loads timezones, and the default timezone for `sas_numeric_to_date()` is now "UTC" instead of "" (#583, @billdenney)
 
 ## New features
 
@@ -33,8 +34,6 @@ These are all minor breaking changes resulting from enhancements and are not exp
 
 * Restyle the package and vignettes according to the [tidyverse style guide](https://style.tidyverse.org) (#548, @olivroy)
 
-* `sas_numeric_to_date()` now warns for timezones other than "UTC" due to potential misinterpretation (#583, @billdenney)
-
 # janitor 2.2.0 (2023-02-02)
 
 ## Breaking changes

R/sas_dates.R

@@ -46,7 +46,7 @@ sas_numeric_to_date <- function(date_num, datetime_num, time_num, tz = "UTC") {
     has_datetime <- TRUE
   }
   if (has_datetime) {
-    ret <- as.POSIXct(datetime_num, origin = "1960-01-01", tz = "UTC")
+    ret <- as.POSIXct(datetime_num, origin = "1960-01-01", tz = tz)
   } else if (has_date) {
     ret <- as.Date(date_num, origin = "1960-01-01")
   } else if (has_time) {

tests/testthat/test-sas_dates.R

@@ -27,7 +27,7 @@ test_that("sas_numeric_to_date", {
   # Timezone warning (#583)
   expect_warning(
     sas_numeric_to_date(date_num = 1, time_num = 1, tz = "America/New_York"),
-    regexp = "`tz` in SAS does not appear to be stored with the source data; please verify timezone conversion is correct",
+    regexp = "SAS may not properly store timezones other than UTC. Consider confirming the accuracy of the resulting data.",
     fixed = TRUE
   )
 })