R Shiny: reactive column name with mutate?

Question

In my reactive dataframe, one column has a reactive name (chosen by the user) and I would like to generate a column whose values are the logarithm of the original column. To do so, I use mutate in the dplyr package. However, when I try to make the name of this new column reactive, there's an error.

For example, in the code below, I name the new column "logarithm" and it works fine:

library(shiny)
library(DT)
library(data.table)
library(dplyr)

ui <- fluidPage(
  titlePanel(""),
  fluidRow(
    checkboxInput(inputId = "logarithm",
                  label = "Log(variable)"),
    dataTableOutput("my_df"),
    textInput("new_name", 
              label = "New_name"),
    actionButton("new_name2", "Validate")
  )
)


server <- function(input, output) {

  data <- head(mtcars[, 1:3])


  reactive_data <- eventReactive(input$new_name2, {
    colnames(data) <- c("mpg", "cyl", input$new_name)
    data
  }) 

  output$my_df <- renderDataTable({
    data <- reactive_data()
      if(input$logarithm){
        data %>%
          mutate(logarithm = log(data[, input$new_name]))

      }
    else {
      data
    }
  })
}

shinyApp(ui = ui, server = server)

But change "logarithm" by "logarithm(input$new_name)" and it won't work anymore.

Does anybody have a solution?

Eli Berkow · Accepted Answer · 2019-09-12T13:21:45.673

1

Based on this question and answer

if(input$logarithm){
        log_name <- paste0('logarithm(', input$new_name, ')')
        data %>%
          mutate(!!log_name := log(data[, input$new_name]))

}

Full code:

library(shiny)
library(DT)
library(data.table)
library(dplyr)

ui <- fluidPage(
  titlePanel(""),
  fluidRow(
    checkboxInput(inputId = "logarithm",
                  label = "Log(variable)"),
    dataTableOutput("my_df"),
    textInput("new_name", 
              label = "New_name"),
    actionButton("new_name2", "Validate")
  )
)


server <- function(input, output) {

  data <- head(mtcars[, 1:3])


  reactive_data <- eventReactive(input$new_name2, {
    colnames(data) <- c("mpg", "cyl", input$new_name)
    data
  }) 

  output$my_df <- renderDataTable({
    data <- reactive_data()
    if(input$logarithm){
            log_name <- paste0('logarithm(', input$new_name, ')')
            data %>%
              mutate(!!log_name := log(data[, input$new_name]))

    }
    else {
      data
    }
  })
}

shinyApp(ui = ui, server = server)

edited Sep 12 '19 at 13:21

answered Sep 12 '19 at 12:13

Eli Berkow

2,628
1
12
22

It does not create the new column in the table, does it work for you? – bretauv Sep 12 '19 at 12:51
yes, make sure you have dplyr version >= 0.7 as per the link. I have added full code above – Eli Berkow Sep 12 '19 at 13:22
It works with the full code you provide (I have dplyr 0.8.3), I must have made a mistake in the original code, thanks a lot – bretauv Sep 12 '19 at 13:58
Sure, glad to help – Eli Berkow Sep 12 '19 at 14:08

R Shiny: reactive column name with mutate?

1 Answers1