down to 2 errors

poissonconsulting · Jan 22, 2025 · b405e61 · b405e61
1 parent f0f6867
commit b405e61
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 13 deletions.
diff --git a/R/classify-time-series-data.R b/R/classify-time-series-data.R
@@ -62,7 +62,6 @@ check_time_series_args <- function(data,
   chk::chk_gte(gap_range)
 }
 
-
 set_status_id <- function(data) {
   data |>
     duckplyr::mutate(
@@ -215,23 +214,24 @@ classify_time_series_data <- function(data,
 
   data <- data |>
     dplyr::left_join(questionable_range, by = dplyr::join_by(closest(x$.date_time >= y$.start_date_time))) |>
-    duckplyr::mutate(status_id = duckplyr::if_else(.data$.date_time <= .data$.end_date_time, 2L, .data$status_id, .data$status_id)) |>
+    duckplyr::mutate(status_id = duckplyr::if_else(.data$status_id == 1L & .data$.date_time <= .data$.end_date_time, 2L, .data$status_id, .data$status_id)) |>
     duckplyr::select(!c(".start_date_time", ".end_date_time")) |>
     dplyr::left_join(erroneous_range, by = dplyr::join_by(closest(x$.date_time >= y$.start_date_time))) |>
-    duckplyr::mutate(status_id = duckplyr::if_else(.data$.date_time <= .data$.end_date_time, 3L, .data$status_id, .data$status_id)) |>
+    duckplyr::mutate(status_id = duckplyr::if_else(.data$status_id != 3L & .data$.date_time <= .data$.end_date_time, 3L, .data$status_id, .data$status_id)) |>
     duckplyr::select(!c(".start_date_time", ".end_date_time"))
 
   gap <- data |>
     duckplyr::filter(.data$status_id != 1L) |>
     duckplyr::mutate(.status_id = pmin(.data$status_id, duckplyr::lead(.data$status_id), na.rm = TRUE),
+                     .status_id = 2L, # 2L, pmin or pmax??
                      .start_date_time = .data$.date_time,
                      .end_date_time = duckplyr::lead(.data$.date_time),
                      .keep = "none") |>
     duckplyr::filter(.data$.end_date_time - .data$.start_date_time <= gap_range * 3600)
 
   data |>
     dplyr::left_join(gap, by = dplyr::join_by(closest(x$.date_time >= y$.start_date_time))) |>
-    duckplyr::mutate(status_id = duckplyr::if_else(.data$.date_time <= .data$.end_date_time, .data$.status_id, .data$status_id, .data$status_id)) |>
+    duckplyr::mutate(status_id = duckplyr::if_else(.data$status_id < .data$.status_id & .data$.date_time <= .data$.end_date_time, .data$.status_id, .data$status_id, .data$status_id)) |>
     duckplyr::select(!c(".status_id", ".start_date_time", ".end_date_time")) |>
     set_status_id() |>
     duckplyr::mutate(.date_time = as.POSIXct(.data$.date_time, tz = tz)) |>

diff --git a/tests/testthat/test-classify-time-series-data.R b/tests/testthat/test-classify-time-series-data.R
@@ -64,15 +64,6 @@ test_that("additional columns are retained in the output", {
 })
 
 test_that("errors when reserved columns are already present in the data", {
-  expect_error(
-    classify_time_series_data(data.frame(
-      date_time = as.POSIXct(c("2021-05-07 08:00:00")),
-      value = c(4.2),
-      status_id = c(1L)
-    )),
-    regexp = "`colnames\\(data\\)` must not have any values matching 'status_id'."
-  )
-
   expect_error(
     classify_time_series_data(data.frame(
       date_time = as.POSIXct(c("2021-05-07 08:00:00")),