Table generated by shiny with incorrect order-CodePudding

This question is similiar to

CodePudding user response：

When you use match, you need to either (a) give it every possible distinct value, or (b) know that a returned NA is inherently what you want/need.

In the use of

arrange(match(WeekP, weekdays(input$daterange1)))

it is matching the values of several values again the values of just two things. You aren't telling it to sort based on a full sequence of things, you are telling R "please sort this vector of things based on two strings". You are assuming that R will infer what those two strings are, and how to project that inference onto the rest of your data.

In this case, since weekdays(input$..) only includes two days of the week, everything else in your match(.) will be NA, which is obviously broken. Had you further debugged your code and looked at the output of just match(.., weekdays(..)) (assign to some column), you would see

    left_join(meanTest, wk_port2eng, by = c("Week" = "WeekE")) %>%      
      select(-WeekP) %>%
      mutate(sortcol = match(Week, weekdays(input$daterange1)))
# # A tibble: 7 x 4
#   Week      Category  mean sortcol
#   <chr>     <chr>    <dbl>   <int>
# 1 Friday    FDE          3      NA
# 2 Monday    FDE          4      NA
# 3 Saturday  FDE          3       2
# 4 Sunday    FDE          4       1
# 5 Thursday  FDE          2      NA
# 6 Tuesday   FDE          6      NA
# 7 Wednesday FDE          3      NA

for which arrange will not magically infer your needs.

Use seq on your input range.

    left_join(meanTest, wk_port2eng, by = c("Week" = "WeekE")) %>%      
      select(-WeekP) %>%
      mutate(sortcol = match(Week, unique(weekdays(seq(input$daterange1[1], input$daterange1[2], by = "day")))))
# # A tibble: 7 x 4
#   Week      Category  mean sortcol
#   <chr>     <chr>    <dbl>   <int>
# 1 Friday    FDE          3       6
# 2 Monday    FDE          4       2
# 3 Saturday  FDE          3       7
# 4 Sunday    FDE          4       1
# 5 Thursday  FDE          2       5
# 6 Tuesday   FDE          6       3
# 7 Wednesday FDE          3       4

So that one reactive block should be:

  data_subset <- reactive({
    req(input$daterange1)
    req(input$daterange1[1] <= input$daterange1[2])
    days <- seq(input$daterange1[1], input$daterange1[2], by = 'day')
    Test1 <- dplyr::filter(data(), date1 %in% days)
    weeks_inp <- unique(weekdays(days))  
    # wk <- wk_port2eng[wk_port2eng$WeekP %in% weeks_inp,]  ###  if weekday is in Portuguese in your notebook
    wk <- wk_port2eng[wk_port2eng$WeekE %in% weeks_inp,]  ###  if weekday is in English in your notebook
    
    weeks_ine <- wk$WeekE
    meanTest1 <- data() %>%
      group_by(Week = tools::toTitleCase(Week), Category) %>% 
      summarise(mean = mean(time, na.rm = TRUE), .groups = 'drop')
    meanTest <- meanTest1[meanTest1$Week %in% as.character(weeks_ine),]
    left_join(meanTest, wk_port2eng, by = c("Week" = "WeekE")) %>%      
      select(-WeekP) %>%
      arrange(match(Week, unique(weekdays(seq(input$daterange1[1], input$daterange1[2], by = "day")))))
  })

Producing:

BTW, this works fine as long as you are always summarizing by day-of-week, in which case weekdays(.) should always return a unique vector.