(v1.7.1.9037) dplyr grouping fix on windows?

This commit is contained in:
dr. M.S. (Matthijs) Berends committed 2021-08-31 17:06:44 +02:00
1 parent 953dfac9e5
commit bcab74fb6d
9 files changed
+38 -49

No files matched your search

+4 -6
View File
@@ -117,14 +117,12 @@ if (AMR:::pkg_is_available("dplyr")) {
first_isolate(col_date = "date",
col_mo = "mo",
col_patient_id = "patient_id",
info = FALSE,
require_cur_data = TRUE),
info = FALSE),
example_isolates %>%
first_isolate(col_date = "date",
col_mo = "mo",
col_patient_id = "patient_id",
info = FALSE,
require_cur_data = TRUE))
info = FALSE))
# support for WHONET
expect_message(example_isolates %>%
@@ -135,8 +133,8 @@ if (AMR:::pkg_is_available("dplyr")) {
first_isolate(info = TRUE))
# groups
x <- example_isolates %>% group_by(ward_icu) %>% mutate(first = first_isolate(require_cur_data = TRUE))
y <- example_isolates %>% group_by(ward_icu) %>% mutate(first = first_isolate(., require_cur_data = TRUE))
x <- example_isolates %>% group_by(ward_icu) %>% mutate(first = first_isolate())
y <- example_isolates %>% group_by(ward_icu) %>% mutate(first = first_isolate(.))
expect_identical(x, y)
}