table subsetting is not happening in shiny reactively when using multiple filtering conditions

Question

I have a dataset :

vru_line    call_id customer_id priority    type    date
AA0101  38080   50395060    2   PS  4/1/2019
AA0101  38080   50395060    0   PS  4/1/2019
AA0101  38081   50353564    2   NW  4/1/2019
AA0102  38082   13000567    2   PS  4/2/2019
AA0102  38083   50395060    1   NW  4/2/2019
AA0102  38084   50353564    0   PS  4/2/2019
AA0103  38085   50353564    1   NW  4/3/2019
AA0103  38086   13000567    1   PS  4/3/2019
AA0103  38087   13000567    0   NW  4/3/2019

Here date is in (mm/dd/yyyy) format.

What I am trying to do using flexdashboard and shiny is to subset the dataset depending upon the conditions:

Filter either for "ALL" vru_line or for Selected one from a drop down menu.
Filter between selected date range

The code below is from the project I am working on:

library(DT)
library(data.table)
library(dplyr)
library(shiny)
library(shinydashboard)
library(tidyverse)
library(stringr)
library(lubridate)
library(anytime)

# this will create a "File Upload" button
fileInput("file1", "Load File:",
          accept = c("text/csv", "text/comma-separated-values, text/plain", ".csv", "text/tab-separated-values", ".tsv") )
checkboxInput("header", "Is first row the Header?", TRUE)

data_set <- reactive({
    req(input$file1)
    inFile <- input$file1
    data_set <- read.csv(inFile$datapath, header=input$header, stringsAsFactors = F)
})


# Date input placeholder for date selection
dateRangeInput("dateRange", "Select Start & End Dates:",
               start = "2019-04-01", end = "2019-04-02",
               separator = " - ", format = "yyyy-mm-dd")

hr()

observe({
    require(dplyr)
    req(input$file1)
    choices = c("ALL", unique(as.character(data_set()$vru_line)))

    updateSelectInput(session,"VarRep",
                             label = "Select VRU",
                             choices = choices, selected=choices[1])

})

selectInput(inputId = "VarRep", label = "Select VRU", 
            multiple = FALSE, choices = list("ALL"), selected = "ALL" )

hr()

# click button
actionButton("displayRes","Display")

# Should select the sebset of data depending on the coditions above and below
eventReactive(input$displayRes, {
  output$VarReptbl <- renderDataTable({ 
    if(input$VarRep == "ALL") {
      data_set() %>% select(everything()) %>%
        filter(anydate(data_set()$date) >= anydate(input$dateRange[1])  &
                 anydate(data_set()$date) <= anydate(input$dateRange[2])) %>% datatable()
    } else {
      data_set() %>% select(everything()) %>%
        filter(as.character(data_set()$vru_line) == input$VarRep) %>%
        filter(anydate(data_set()$date) >= anydate(input$dateRange[1])  &
                 anydate(data_set()$date) <= anydate(input$dateRange[2])) %>% datatable()
    }
  })
})


hr()
dataTableOutput("VarReptbl")

There is no error in the code being displayed and when I import the data and click on the display button, no table is returned.

Not sure where I am doing wrong? The code looks ok though.

What is expected that if I select one of the options from drop down menu and pick two dates then the data should be filtered as a subset and displayed as the datatable.

Have you tried placing `browser()` in a line that is suppose to generate the table in question, or perhaps tried printing it to the console? What does it return? — Roman Luštrik, Aug 28 '19 at 11:45
it returns a table. very much similar to the example in `https://community.rstudio.com/t/reactive-date-range-input-filtering-data-set-with-observe-events/20313` — loveR, Aug 28 '19 at 11:46
What I would try is use `eventReactive` to output a data.frame and pass that into `output$VarReptbl`. — Roman Luštrik, Aug 28 '19 at 11:54
If I don't apply date filter, I am able to get the result in some other example within the project, but in this case, the above seems to have some issues related to date based filtering and if-else loop. not sure where I am wrong and stuck up. — loveR, Aug 28 '19 at 11:58

score 2 · Answer 1 · answered Aug 28 '19 at 13:41

2

Output bindings inside any reactive is not good practice and almost never needed. Here's what you probably need. -

some_data <- eventReactive(input$displayRes, {
  data_set() %>%
    filter(as.character(vru_line) == input$VarRep | input$VarRep == "ALL") %>%
    filter(anydate(date) >= anydate(input$dateRange[1])  &
             anydate(date) <= anydate(input$dateRange[2]))
  }
})

output$VarReptbl <- renderDataTable({
  some_data() %>% datatable()
})

answered Aug 28 '19 at 13:41

Shree

10,835
1
14
36

when I select "ALL", I get no result printed (meaning, no rows/observations displayed) but empty datatable is shown, and when I select `VRU as AA0101` or any other, I get error `Result must have length 100, not 500` – loveR Aug 28 '19 at 13:57
@loveR The issue is probably related to date filters. Just make sure `anydate()` is working as expected. – Shree Aug 29 '19 at 13:42
I think anydate is not working properly in my case. so I used lubridate package function mdy() – loveR Aug 29 '19 at 13:53
@loveR In that case just replace `anydate` with `mdy` above and it should work fine. – Shree Aug 29 '19 at 13:54

score 0 · Answer 2 · answered Aug 29 '19 at 12:03

Although lazy, and not lean, but this solution seems to be working for me (Thanks to Shree above):

some_data <- eventReactive(input$displayRes, {
  testD <- subset(data_set(), vru_line= =input$VarRep | input$VarRep == "ALL") %>%
    select(vru_line, date, text)
  testD$date <- lubridate::mdy(testD$date)
  testD1 <- subset(testD, date >= input$dateRange[1]  &
                    date <= input$dateRange[2])
  return(testD1)
})

table subsetting is not happening in shiny reactively when using multiple filtering conditions

2 Answers2