CODEWITHSUNDEEP
rshinyplotly
daniell
daniell

Reputation: 179

Bar chart failing to render (Error in : First argument, `data`, must be a data frame or shared data. [No stack trace available])

I have the following data:

> data
creadate     employee  
2020-01-11        A A
2020-01-11        A A
2020-01-11        A A   
2020-01-12        A B 
2020-01-12        A B      
2020-01-12        A C      
2020-01-13        A C      
2020-01-14        A C      
2020-01-14        A D      
2020-01-15        A D

The objective is to build a shiny app containing a plotly graph. First, I want to transform the data and I add another column that counts the number of occurrences for each time frame(week, day, and date)-employee combination. E.g. the normal code for the week would be:

data %>% 
  count(Week = week(creadate), employee)

Now, I want to give the user the possibility to choose between the three choices. I write the following full code (see #2 specifically)

library(shiny)
library(shinyWidgets)
library(tidyverse)
library(DT)
library(plotly)

data <- read.csv("Desktop/adf.csv")

ui <- fluidPage(
  fluidRow(
    column(4, 
           selectInput("timeFrame", "Select time frame", c("month", "week", "day"), multiple = TRUE), 
           dateRangeInput("date", "Select date", start  = "2020-01-01"),
    column(8, 
           plotlyOutput("errorTable")
    )
  )
)

server <- function(input, output, session) {  
  
  #create a reactive object with a NULL starting value
  listofrows <- reactiveValues(data = NULL)
  
  #observe the changes in inputs and update the reactive object 
  observeEvent(c(input$timeFrame, input$date), {
    
    listofrows$data <-  data %>% 
      subset(data, creadate >= input$date[1] & creadate <= input$date[2]) %>% # 1------------------------
      count(input$timeFrame[1] = input$timeFrame[1](creadate), employee) # 2-----------------------------
      

  }, ignoreInit = T,ignoreNULL = TRUE)
   
  
  output$plot <- renderPlotly({
    plot_ly(new_data,
      x = ~ input$timeFrame[1], 
      y = ~ n, 
      color = ~ employee, 
      frame = ~ input$timeFrame[1], 
      type = 'scatter',
      mode = 'markers'
    )
  }
  ) 
 }

shinyApp(ui, server)

However, this code does not work. It tells:

Error in : First argument, `data`, must be a data frame or shared data.
  [No stack trace available]```
Do you have any idea?

Upvotes: 0

Views: 155

Answers (1)

Konrad Rudolph
Konrad Rudolph

Reputation: 545528

There are multiple problems.

The first is with the subset call, as indicated by the warning: you are passing data twice — once via the first explicit argument, and once since you’re using subset on the right-hand side of a pipe; the code

data %>% 
  subset(data, creadate >= input$date[1] & creadate <= input$date[2]) …

is the same as

subset(data, data, creadate >= input$date[1] & creadate <= input$date[2]) …

To fix this, remove the data argument. Furthermore, since you’re using the tidyverse, I recommend sticking to it, rather than mixing it with base functions; in other words, use filter instead of subset. Besides consistency, the filter function is also more robust in the face of errors and has better error messages than subset.

the second problem is your count function call — it’s not entirely clear what you’re attempting to do here. Consult the documentation of the count function to ensure you’re using it correctly.

There are more issues. For example, your plotting code uses a new_data variable that’s never defined. You probably want to use listofrows$data here. The arguments to plot_ly are also wrong: you can’t pass variables as formulas, you need to quote them properly (e.g. via formula(paste('~', input$timeFrame[1]))).

Upvotes: 1

Related Questions