put back datawizard::data_unite and use alternative trick to match order

R/report_sample.R

@@ -177,8 +177,20 @@ report_sample <- function(data,
     # for more than one group, fix column names. we don't want "a.b (n=10)",
     # but rather ""a, b (n=10)""
     if (length(by) > 1) {
-      group_names <- gsub("[.]", ", ", names(result))
-      result <- stats::setNames(result, group_names)
+      old_names <- datawizard::data_unite(
+        unique(groups),
+        new_column = ".old_names",
+        separator = "."
+      )[".old_names"]
+
+      new_names <- datawizard::data_unite(
+        unique(groups),
+        new_column = ".new_names",
+        separator = ", "
+      )[".new_names"]
+
+      name_map <- stats::setNames(new_names[[1]], old_names[[1]])
+      names(result) <- name_map[names(result)]
     }
     # remember values of first columns
     variable <- result[[1]]["Variable"]