CODEWITHSUNDEEP
rdataframeshinydplyrtidyr
Syrah.Sharpe
Syrah.Sharpe

Reputation: 103

Complete() function from dplyr / tidyverse does not work with Shiny interactive variables

I am trying to fix the following error: Warning: Error in : Please supply variables to complete Stack trace (innermost first)

While building a ShinyApp to graph my data through bar charts, first I wanted to check that the Shiny interactives would work in my code. However, the function complete_() does not seem to work with shiny interactives. I created the following ui and server to test this: My data:

gender <- c("Male", "Female")
residency <-c("InsideUS", "OutsideUS")
category <- c("A", "B")
SES <- c("Lower", "Middle", "Upper")
choices <- c("gender", "residency", "SES")
variables <- c("SES")
cat<- c("category")
df <- data.frame( gender=as.factor(sample(gender, size=100, replace=TRUE)), 
residency=as.factor(sample(residency, size=100, replace=TRUE)), 
category=as.factor(sample(category, size=100, replace=TRUE)), 
SES=as.factor(sample(SES, size=100, replace=TRUE)))

My ui:

 ui <- fluidPage(
      sidebarLayout(
    sidebarPanel(
      selectInput(inputId = "choiceDisplay1",
                  label= "Please select an option",
                  choices= c("Gender"="gender",
                            "Residency"="residency",
                            "Social Economic Status"="SES"),
                  selected="gender")
     ),
      mainPanel(
      verbatimTextOutput("summary")
    )
  )
)

My server:

server <- function(input, output) {
#Target Market
  output$summary <- renderPrint({
    df %>% group_by_(.choices=input$choiceDisplay1, .cat=c("category"), 
.variables=c("SES")) %>%
            summarize(n=n()) %>%
            ungroup() %>%
            complete_(.choices=input$choiceDisplay1, .cat=c("category"), 
 .variables=c("SES"),
             fill=list(n=0)) %>%
             group_by_(.choices=input$choiceDisplay1, .cat=c("category")) %>%
            mutate(perc= n/sum(n))
    })
}

I know the problem first arises with complete_(). Any insights on to what is causing this?

Thank you!

Upvotes: 3

Views: 1160

Answers (1)

Florian
Florian

Reputation: 25385

The function works fine with Shiny reactive variables, but there are some small issues with your code. First, when we type help(complete_), we read that "the underscored versions are now superfluous", so we can use the regular ones. Furthermore, in the complete and the subsequent group_by statements, we only have to refer to the columns, So instead of 

.choices=input$choiceDisplay1, .cat=c("category"), .variables=c("SES")

we can just do 

.choices, .cat, .variables,

since those are the columns we want to complete. A working example is given below. Note that I have limited your dataset to 20 records so we can actually see the missing records being added by complete.

Hope this helps!

gender <- c("Male", "Female")
residency <-c("InsideUS", "OutsideUS")
category <- c("A", "B")
SES <- c("Lower", "Middle", "Upper")
choices <- c("gender", "residency", "SES")
variables <- c("SES")
cat<- c("category")

set.seed(1)
df <- data.frame( gender=as.factor(sample(gender, size=20, replace=TRUE)), 
                  residency=as.factor(sample(residency, size=20, replace=TRUE)), 
                  category=as.factor(sample(category, size=20, replace=TRUE)), 
                  SES=as.factor(sample(SES, size=20, replace=TRUE)))

library(dplyr)
library(tidyr)
library(shiny)

ui <- fluidPage(
  sidebarLayout(
    sidebarPanel(
      selectInput(inputId = "choiceDisplay1",
                  label= "Please select an option",
                  choices= c("Gender"="gender",
                             "Residency"="residency",
                             "Social Economic Status"="SES"),
                  selected="gender")
    ),
    mainPanel(
      verbatimTextOutput("summary")
    )
  )
)

server <- function(input, output) {
  #Target Market
  output$summary <- renderPrint({
    df %>% group_by_(.choices=input$choiceDisplay1, .cat=c("category"), 
                           .variables=c("SES")) %>%
      summarize(n=n()) %>%
      ungroup()  %>% complete(.choices, .cat, .variables,
                              fill=list(n=0)) %>%
      group_by(.choices, .cat) %>%
      mutate(perc= n/sum(n))
  })
}

shinyApp(ui,server)

Upvotes: 1

Related Questions