Error : invalid 'envir' argument of type 'character'

[Harshad Patil]

I am getting error "invalid 'envir' argument of type 'character'" while executing shiny R code. Please help as I am unable to figure out the workout.

  library(MASS)
  library(shinythemes)
  library(shiny)
  library(ggplot2)
 
  mass.tmp <- data(package = "MASS")[3]
  mass.datasets <- as.vector(mass.tmp$results[,3])
 
  ui <- fluidPage(
   
    theme = shinytheme("superhero"),
    titlePanel("Linear Regression Modelling"),
    sidebarLayout(
      sidebarPanel(
        selectInput("dsname", "Dataset:",choices = c(mass.datasets)),
        uiOutput("y_axis"),
        uiOutput("x_axis")
      )     ,
      mainPanel(
        tags$br(),
        tags$br(),
        "R-squared:",
        tags$span(tags$b(textOutput("rsquared")),style="color:blue")
      )
     
    )
  )
  server <- function(input, output) {
   
    output$x_axis <- renderUI({
      col_opts <- get(input$dsname)
      selectInput("x_axis2", "Independent Variable:", choices = c(names(col_opts)))
    })
   
    cols2 <- reactive({
      col_opts2 <- get(input$dsname)
      #names(col_opts2)[!grepl(input$x_axis2, names(col_opts2))]
    })
   
    output$y_axis <- renderUI({
      selectInput("y_axis2", "Dependent Variable:", choices = c(names(cols2())))
    })
   
    model <- reactive({
      #lm(input$dsname[,names(input$dsname) %in% input$y_axis2] ~ input$dsname[,names(input$dsname) %in% input$x_axis2])


      lm( as.formula(paste(input$y_axis2," ~ ", input$x_axis2)) , data = input$dsname )
    })
   
    model_summary <- reactive({summary(model())})
    output$rsquared <- renderText({  model_summary()$r.squared   })
 
  }
 
  shinyApp(ui = ui, server = server)

[Bárbara Borges]

You're passing the data argument to lm as a string instead of a variable. I.e., on startup, for example, input$dsname is equal to "Aids2" (a character vector), not  Aids2 (a data.frame).

If you change  data = input$dsname to  data = eval(parse(text = input$dsname)) , it should work fine.

[Harshad Patil]

Thanks. Will check and confirm.